- ABAP SQL
- Introduction
- Excursion: Database Tables and Views
- Reading Data Using SELECT
- Changing Data in Database Tables
- More Information
- Executable Example
-
ABAP SQL is a subset of SQL which is the standardized language for accessing databases.
-
The main ABAP SQL keywords to read and change data are the following:
Keyword Purpose SELECTReads data from database tables INSERTAdds rows to database tables UPDATEChanges the content of rows of database tables MODIFYInserts rows into database tables or changes the content of existing rows DELETEDeletes rows from database tables -
ABAP SQL statements use the ABAP SQL interface. This interface transforms all ABAP SQL statements that access the standard database of an AS ABAP to platform-dependent SQL and forwards the results to the database system.
-
Generally bear in mind the performance notes when using ABAP SQL. The considerations there are not relevant for this cheat sheet since the focus is on syntax options.
Expand to view the details
This section provides bullet points on database tables and views which contain persisted data. Note that the code snippets in this cheat sheet focus on database tables as data source for ABAP SQL statements.
Database tables in AS ABAP ...
- are objects of the ABAP Dictionary (DDIC). The term database table describes a physical database table in the current standard database.
- are two-dimensional matrices consisting of rows and columns.
- contain a table key, i. e. a field or a combination of fields uniquely identifies every row in a table. A primary key must exist for every database table.
- Note the concept of foreign keys in which one or more columns of a database table can be primary keys of another table. See more information here.
- have a flat structure type. Plus, the definition of database tables consists of technical and semantic properties.
- can be referenced as a data type and can be accessed using ABAP SQL.
Find more information in the respective (sub)topics in the ABAP Keyword Documentation here.
Views ...
- are further ABAP Dictionary objects for grouping columns from one or more database tables, among others.
- usually realize a join with defined join conditions.
- Note:
- Similar to database tables, the columns of such a view form a flat structure. The view's name can be used, for example, as data types to declare data objects, too.
- The views can be accessed by ABAP SQL, especially for reading
purposes using
SELECT.
"Classic" DDIC Views ...
- are the oldest form of views and are not available in ABAP Cloud.
- can be accessed by ABAP SQL for read and write operations, however, writing is only supported if the view is created with only one database table.
- can only be created in the ABAP Workbench.
"Modern" Views (since release 7.40)
- External
views
as proxies for SAP HANA
views
(attribute view, analytic view, calculation view)
- SAP HANA Views are entities of the SAP HANA database that are defined using the SAP HANA Studio.
- They are based on HANA-specific data types.
- Using external views of the ABAP dictionary, you can make those SAP HANA views "known" to the ABAP program. In doing so, the external views can be used like classic DDIC views as structured data types and as a source for reading operations with ABAP SQL.
- To be used only if the central database of the AS ABAP is an SAP HANA database.
- ABAP Core Data Services (ABAP
CDS)
...
- serve the purpose of defining semantically rich data models.
- have a lot more options than classic views, for example, they support annotations. Data sources can be combined using associations, views can be defined with input parameters, and more.
- are used like a classic database view as structured data types
and used as a source for reading operations with ABAP SQL (using
SELECT). - are created using Data Definition Language (DDL) in the ADT (that is, a source code editor, in contrast to a form-based editor).
- are, in contrast to external views, supported by all database systems (that support the ABAP CDS characteristics).
You use ABAP SQL SELECT statements to read data from one or more database tables (or views). This can be done to create a multirow or single row result set by assigning the result set to a suitable data object, i. e. you can store the multirow read result in an internal table or the single row result in a structure.
The SELECT statement includes several clauses that serve
different purposes. The following code snippet shows the basic syntax (see the note below for a different but interchangeable syntax):
SELECT FROM source "What database table or view to read from
FIELDS field_list "What columns should be read
WHERE condition "Specifies conditions on which a row/rows should be read
INTO @target. "Data object to which the result set is assigned (preceded by @)💡 Note
- There are further clauses available of which some are dealt with further down. In general, the recommendation is to hit
F1for the keywords and additions to get all the details in the ABAP Keyword Documentation.- Especially in older ABAP programs, you will see other forms of the
SELECTsyntax that you should no longer use. Strict syntax check modes might enforce the use of specific ABAP SQL syntax. For example, theINTOclause should be placed after the other clauses. Furthermore, host variables or host expressions are required for data objects and expressions, i. e. they must be preceded by@or@( ... ). Further information: Release-Dependent Syntax Check Modes (F1 docu for standard ABAP).- Regarding host variables as in
SELECT ... INTO @target.and since they are used in most examples below: A host variable is a data object that is
- declared in the ABAP program
- prefixed with the
@character and- specified in an operand position of an ABAP SQL statement. See more information here.
- The
SELECTlist, i. e. the fields that are specified, can also be specified following theSELECTkeyword before theFROMclause - withoutFIELDS. The following twoSELECTstatements are basically the same but differently arranged. The code snippets in the cheat sheet randomly use one syntax or the other.SELECT FROM dbtab FIELDS comp1, comp2, comp3 ... SELECT comp1, comp2, comp3 FROM dbtab ...- Regarding the target into which data is read: Instead of using a variable that is (extra) declared beforehand, you can also make use of inline declarations, for example
... INTO TABLE @DATA(itab)., to comfortably create an appropriate variable in place. Note that in case of internal tables as targets, the resulting table is a standard table and has an empty key which might have an impact when further processing the internal table entries. Find more information in the ABAP cheat sheet Internal Tables. The declaration operatorFINALcan be used to declare immutable variables.- Many syntax examples in this cheat sheet show a selection from
dbtabdenoting a database table as a source. However, other data sources can also be specified.
Reading a single row into a structure
"SINGLE addition
"Here, all fields of a single row a read. Specifying an
"asterisk * indicates that all fields are to be read.
"Alternatively, you can list all the fields separated by comma.
"Note that if the selection covers more than one row, e. g. in case
"of a non-unique WHERE clause, one of these rows is included in
"the result.
SELECT SINGLE FROM dbtab
FIELDS *
WHERE ...
INTO @struc. "Existing structure of dbtab's row type
"Reading a selected set of fields of a single row
SELECT SINGLE FROM dbtab
FIELDS comp1, comp2, comp3
WHERE ...
INTO @DATA(struc2). "Structure declared inline
"Alternative syntax without the FIELDS addition
"Here, the CORRESPONDING FIELDS OF addition is used. Only the content of
"columns that have identically named components in the target data object
"is assigned.
SELECT SINGLE comp1, comp2, comp3 "Selected set of fields
FROM dbtab
WHERE ...
INTO CORRESPONDING FIELDS OF @struc. "Existing structure💡 Note
- Although its use is optional, a
WHEREclause should be specified to further restrict the read result.- Regarding the addition
CORRESPONDING FIELDS OFin theINTOclause: As mentioned, only the content of columns for which there are identically named components in the target are assigned. However, if you want to read data into an existing data object and particular fields are specified in theSELECTlist and if the addition is not specified, you might stumble on undesired results. The target data object must contain enough components and the content of the columns are assigned to the components of the target from left to right in the order specified afterSELECT. The content of surplus components of the target is not changed. Plus, pay attention to assignment rules. Basic rule: WithoutCORRESPONDING ..., column names do not play a role but only the position. WithCORRESPONDING ..., the position of the columns does not play a role but only the name.- Find more information regarding the addition
INTOhere.
Reading multiple rows into an internal table.
SELECT FROM dbtab
FIELDS * "All fields
WHERE ...
INTO TABLE @itab. "itab has an appropriate row type
"Alternative syntax without the FIELDS addition
SELECT comp1, comp2, comp3 "Selected set of fields
FROM dbtab
WHERE ...
INTO TABLE @DATA(lv_itab). "Internal table declared inline
"Selected set of fields, existing variable
"See the note on CORRESPONDING FIELDS OF above
SELECT FROM dbtab
FIELDS comp1, comp2, comp3 "Selected set of fields
WHERE ...
INTO CORRESPONDING FIELDS OF TABLE @itab.SELECT loop: Sequentially reading multiple rows.
- A
SELECTloop can be opened if the assignment is made to a structure and the additionSINGLEis not used. - If the row is found, the system field
sy-subrcis set to0. - The loop must be closed using
ENDSELECT. - To terminate the loop completely, you can use the statement
EXIT. - Note: As covered further down, when using the addition
PACKAGE SIZEand storing the result in a table, a loop is opened, too.
SELECT FROM dbtab
FIELDS *
WHERE ...
INTO @struc.
IF sy-subrc = 0.
... "For example, making changes on data and adding the row to an internal table.
ENDIF.
ENDSELECT.SELECT/FROM clauses:
Checking the existence of a row in a database table
"Instead of @abap_true, you could also use 'X'.
SELECT SINGLE @abap_true
FROM dbtab
WHERE ...
INTO @DATA(exists).
IF exists = abap_true.
...
ENDIF.Removing rows that occur more than once in a multirow result set using the DISTINCT addition.
- Cannot be used with the addition
SINGLE. - See more information here here.
SELECT DISTINCT comp1
FROM dbtab
WHERE ...
INTO TABLE @itab.SELECT list variants (some of them are already outlined above)
The following specifications can also be combined:
SELECT * ...: As outlined above, the*character defines all columns to be read from a data source (in the order specified there).SELECT col1, col2, col3 ...: A comma-separated list of individual column names.SELECT data_source~col1, data_source~col2, data_source~col3 ...: A comma-separated list of individual column names. Here, the name of the data source is explicitly specified and precedes the column name, separated by a tilde.SELECT data_source~* ...: In this case, the name of the data source is followed by a tilde and the*character to specify all columns. Note that there are special conditions when using this variant.SELECT col1 AS al1, col2 AS al2, col3 AS al3 ...:- Defining alias names for individual columns of the result set with
AS. - Make sure that you use an alias name only once here. In the statement, the alias name can only be used after an
ORDER BYclause. - As shown further down, in some cases (e. g. when using SQL expressions) the specification of an alias name is required. Setting an alias name for the data source is also possible (
SELECT FROM dbtab AS alias_name ...). See the section on joins further down.
- Defining alias names for individual columns of the result set with
💡 Note
You have plenty of options regarding the specification of the columns in theSELECTlist, among them, the outlined direct specification of the column name. SQL expressions can be specified, too. See more details here and in the sections on SQL expressions further down.
"All fields
SELECT * FROM dbtab
WHERE ...
INTO ...
"Comma-separated list
SELECT col1, col2, col3
FROM dbtab
WHERE ...
INTO ...
"Comma-separated list, data source explicitly specified
SELECT dbtab~col1, dbtab~col2, col3
FROM dbtab
WHERE ...
INTO ...
"Data source explicitly specified, all fields
SELECT dbtab~*
FROM dbtab
WHERE ...
INTO ...
"Alias names
"Consider the following: You want to read data from a database table into a target data
"object but, for example, a name in the target is different. Provided that there will
"not be an issue regarding the type (conversion) when the values are assigned, you might
"specify an alias name for the database column to match a component's name in the target data object.
SELECT FROM dbtab
FIELDS comp1 AS comp_a, comp2 AS comp_b, comp3 AS comp_c
WHERE ...
INTO CORRESPONDING FIELDS OF TABLE @itab.
"Alias name also possible for the data source
SELECT ds~col1, ds~col2, ds~col3
FROM dbtab AS ds
WHERE ...
INTO ...Reading data from a database table in another client (classic ABAP only). Note that there are several variants of the USING ... addition for switching the implicit client handling (F1 docu for standard ABAP) from the current client to other clients. See more information here (F1 docu for standard ABAP).
"Some examples; not available in ABAP for Cloud Development
"Replaces the current client with the specified client
SELECT *
FROM dbtab USING CLIENT '000'
WHERE ...
INTO TABLE @itab.
"Selects data of any number of clients
SELECT *
FROM dbtab USING ALL CLIENTS
WHERE ...
INTO TABLE @itab.Reading data from an internal table as data source using SELECT. Note that an alias name must be specified for the internal table used as data source.
Find more information here.
SELECT *
FROM @itab1 AS tab
WHERE ...
INTO TABLE @DATA(itab2).INTO clause:
Limiting the number of returned table rows using the optional addition UP TO n ROWS.
"A maximum of five rows are to be returned
"If the INTO clause is the last clause, the UP TO clause must be positioned after it.
SELECT * FROM dbtab
WHERE ...
INTO TABLE @DATA(itab_upto)
UP TO 5 ROWS.Returning only the table rows after a row with a specified count from the result set using the optional addition OFFSET n. You can only use the addition, if an ORDER BY clause is specified.
"In the example, data of all flights are retrieved, except for the 2 flights
"with the shortest flight time.
SELECT *
FROM ztest_abap_flsch
WHERE carrid = 'LH'
ORDER BY fltime ASCENDING
INTO TABLE @DATA(itab)
OFFSET 2.Reading into individual elementary data objects.
Apart from reading into structures and internal tables outlined above, you can also read into individual elementary data objects.
Here, the individual elementary data objects as target objects are specified in a comma-separated list (e. g. as existing host variables or declared inline with @DATA(...)) and put between a pair of parentheses.
Note:
- The comma-separated list must have the same number of elements as columns in the result set.
- The content of the columns in the result set is assigned to the data objects specified in the list from left to right in accordance with the order specified in the
SELECTlist. - Note the assignment rules also in this context.
- More information here.
SELECT FROM dbtab
FIELDS comp1, comp2, comp3
WHERE ...
INTO (@res1,@res2,@res3).
"INTO (@DATA(res1),@DATA(res2),@DATA(res3)). "Using inline declarationsAppending the result set to an existing internal table.
The addition INTO initializes the target object. When using the addition APPENDING, you can retain existing lines in internal tables. APPENDING is also possible with the addition CORRESPONDING FIELDS OF TABLE.
SELECT * FROM dbtab
WHERE ...
APPENDING TABLE @itab.
SELECT * FROM dbtab
WHERE ...
APPENDING CORRESPONDING FIELDS OF TABLE @diff_itab.Reading into packages of a specified number of rows when reading into internal tables. The addition PACKAGE SIZE n can be specified after INTO TABLE and APPENDING TABLE. A SELECT loop ist opened. After PACKAGE SIZE, the number of rows is specified (which can be a host variable, host expression or a literal of type i) denoting the number of rows to be inserted in the target object per iteration.
SELECT FROM dbtab
FIELDS comp1, comp2, comp3
WHERE ...
INTO TABLE @DATA(itab_pack) PACKAGE SIZE n.
...
ENDSELECT.Specifying an anonymous data object as target object using the addition NEW. Only to be used after INTO and not APPENDING.
"Here, the target object is an anonymous data object declared inline.
SELECT FROM dbtab
FIELDS comp1, comp2, comp3
WHERE ...
INTO TABLE NEW @DATA(dref).GROUP BY
clause: Combining groups of table rows in the result set. You
might also use SQL expressions
here. Multiple clause elements are separated by a comma. Find more
information on SQL expressions further down.
Note that the GROUP BY clause requires all columns that are
directly specified in the SELECT list or specified there as an
argument of an SQL expression to be specified. An exception to this is
aggregate
functions
in aggregate
expressions
(except grouping
functions)
as shown in the following example.
In the example below, the database table rows that have the same content in column comp1 are combined. The lowest and highest values in column comp2 are determined for each of these groups and placed into the combined row.
SELECT FROM dbtab
FIELDS comp1, MIN( comp2 ) AS min, MAX( comp2 ) AS max
WHERE ...
GROUP BY comp1
INTO ...HAVING
clause: Limiting the number of table rows in groups in the
result by setting conditions on these rows. The rows for which a
logical expression is true are inserted in the target variable. Note
that HAVING can only be used together with GROUP BY.
SELECT FROM dbtab
FIELDS comp1, MIN( comp2 ) AS min, MAX( comp3 ) AS max
WHERE ...
GROUP BY comp1
HAVING comp1 LIKE '%XYZ%' AND SUM( comp4 ) > 100
INTO ...ORDER BY
clause: Sorting the result set by specified columns.
The following example shows the ordering of the result set based on the content of the primary key of the data source. You can also order by any columns and by explicitly specifying the sort order. There are more ordering options, for example, by using SQL expressions.
SELECT FROM dbtab
FIELDS comp1, comp2, comp3
WHERE ...
ORDER BY PRIMARY KEY
"comp2 ASCENDING
"comp2 DESCENDING
INTO ...💡 Note
- Not specifying
ORDER BYmeans that the order of entries in the result set is undefined.- If
ORDER BYandGROUP BYclauses are used, all columns specified afterORDER BYmust also be specified afterGROUP BY.- If aggregate functions are specified after
SELECT, all columns that are specified afterORDER BYand that do not have an alias name for an aggregate function must also be specified afterSELECTand after theGROUP BYclause which is required in this case, too.
WHERE clause: Restricts the number of rows that are included in the result set using logical expressions. See further information on them in the following sections.
SELECT FROM dbtab
FIELDS comp1, comp2, comp3
WHERE comp1 = 'abc'
AND comp2 < 123
INTO ...ABAP offers plenty of SQL
operands
and
expressions
that are possible in ABAP SQL statements, not only in the context of
SELECT statements and the SELECT lists which are
mainly used for the following demonstration examples. Questions about
when to use what, what is possible in which contexts and positions, is
beyond the scope of this cheat sheet. Check the details in the
respective topics in the ABAP Keyword Documentation. Find a general
overview of important operand positions in ABAP SQL
here.
Due to the rich variety of options, the cheat sheet covers a selection.
- Are elementary operands in an ABAP SQL statement
- Can be database table or view columns, a
literal,
host
variables
(i. e. global or local data objects escaped using
@:@dobj) or host expressions (@( ... ))- Regarding literals: They are not prefixed with the escape
character
@. The literals can be typed (using the type name and content within a pair of backquotes:char`abc`) with built-in ABAP Dictionary types or untyped. Typed literals are preferable for the following reasons: Using untyped literals means extra cost in terms of performance since they must be converted by the compiler. Plus, their use can result in errors at runtime whereas typed literals guarantee type compatibility at once. For more information on typed literals, refer to the ABAP Keyword Documentation and the Typed Literals in ABAP SQL section of the Data Types and Data Objects cheat sheet. - Regarding host expressions: Structures and internal tables are possible as host expressions for statements modifying the content of database tables as shown further down.
- Regarding literals: They are not prefixed with the escape
character
- See more information here.
Example demonstrating possible operands:
DATA upto TYPE i VALUE 3.
SELECT FROM zdemo_abap_flsch
FIELDS
"Specifies a column of a data source directly using its name
cityfrom,
"Column selector ~ can be used to prefix every specified column.
"Here, it is optional. It is non-optional, e. g., if multiple data
"sources in an ABAP SQL statement are edited and the column name
"is not unique.
zdemo_abap_flsch~cityto,
'Lufthansa' AS name, "Untyped literal
char`X` AS flag, "Typed literal
@upto AS num, "Host variable
@( cl_abap_context_info=>get_system_date( ) ) as date "Host expression
WHERE carrid = 'LH' "Untyped literal
AND countryfr = char`DE` "Typed literal
"Data object created inline and escaped with @
INTO TABLE @DATA(it)
"The following clause shows all options having the same effect
UP TO 3 ROWS. "Untyped numeric literal
"UP TO int4`3` ROWS. "Typed numeric literal
"UP TO @upto ROWS. "Host variable
"UP TO @( 10 - 7 ) ROWS. "Host expression- Expressions in an ABAP SQL statement that are passed to the database system for evaluation.
- For example, SQL expressions can be specified as columns in the
SELECTlist as demonstrated in most of the following examples. Find information on more possible positions and general information on SQL expressions here and the subtopics there.
- An elementary expression represents one of the four mentioned operands above: A value from the database (the column name) or values from an ABAP program passed to the database (literal, host variable or host expression).
- As an example, see the
SELECTlist in the example above. - See more information here.
- You can use built-in functions in ABAP SQL.
- Result: Value with the associated dictionary type.
- Arguments of the functions: Cover one or more SQL expressions.
- See more information here.
Example: Numeric functions
SELECT SINGLE
carrname,
"Division, result rounded to an integer
"Result: 2
div( 4, 2 ) AS div,
"Division, 3rd argument: result is rounded to the specified
"number of decimals
"Result: 0.33
division( 1, 3, 2 ) AS division,
"Result is rounded to first greater integer
"Result: 2
ceil( decfloat34`1.333` ) AS ceil,
"Result is the remainder of division
"Result: 1
mod( 3, 2 ) AS mod,
"Result: Largest integer value not greater than the specified value
"Result: 1
floor( decfloat34`1.333` ) AS floor,
"Returns the absolute number
"Result: 2
abs( int4`-2` ) AS abs,
"Result is rounded to the specified position after the decimal separator
"Result: 1.34
round( decfloat34`1.337`, 2 ) AS round
FROM zdemo_abap_carr
WHERE carrid = 'AA'
INTO @DATA(numeric_functions).Example: String functions
SELECT SINGLE
carrid, "LH
carrname, "Lufthansa
url, "http://www.lufthansa.com
"Concatenates strings, ignores trailing blanks
"Result: LHLufthansa
concat( carrid, carrname ) AS concat,
"Concatenates strings, number denotes the blanks that are inserted
"Result: LH Lufthansa
concat_with_space( carrid, carrname, 1 ) AS concat_with_space,
"First letter of a word -> uppercase, all other letters -> lowercase;
"note that a space and other special characters means a new word.
"Result: Http://Www.Lufthansa.Com
initcap( url ) AS initcap,
"Position of the first occurrence of the substring specified
"Result: 6
instr( carrname,'a' ) AS instr,
"String of length n starting from the left of an expression;
"trailing blanks are ignored
"Result: Luft
left( carrname, 4 ) AS left,
"Number of characters in an expression, trailing blanks are ignored
"Result: 24
length( url ) AS length,
"Checks if expression contains a PCRE expression;
"case-sensitive by default (case_sensitive parameter can be specified)
"Notes on the result: 1 = found, 0 = not found
"Result: 1
like_regexpr( pcre = '\..', "Period that is followed by any character
value = url ) AS like_regex,
"Returns position of a substring in an expression,
"3rd parameter = specifies offset (optional)
"4th parameter = determines the number of occurrences (optional)
"Result: 9
locate( carrname, 'a', 0, 2 ) AS locate,
"Searches a PCRE pattern, returns offset of match;
"many optional parameters: occurrence, case_sensitive, start, group
"Result: 21
locate_regexpr( pcre = '\..', "Period followed by any character
value = url,
occurrence = 2 ) "2nd occurrence in the string
AS locate_regexpr,
"Searches a PCRE pattern, returns offset of match + 1;
"many optional parameters: occurrence, case_sensitive, start, group
"Result: 2
locate_regexpr_after( pcre = '.', "Any character
value = url,
occurrence = 1 ) AS locate_regexpr_after,
"Removes leading characters as specified in the 2nd argument,
"trailing blanks are removed
"Result: ufthansa
ltrim( carrname, 'L' ) AS ltrim,
"Counts all occurrences of found PCRE patterns
"Result: 2
occurrences_regexpr( pcre = '\..', "Period that is followed by any character
value = url ) AS occ_regex,
"Replaces the 2nd argument with the 3rd in an expression
"Result: Lufth#ns#
replace( carrname, 'a', '#' ) AS replace,
"Replaces a found PCRE expression;
"more parameters possible: occurrence, case_sensitive, start
"Result: http://www#ufthansa#om
replace_regexpr( pcre = '\..', "Period that is followed by any character
value = url,
with = '#' ) AS replace_regex,
"Extracts a string with the length specified starting from the right
"Result: hansa
right( carrname, 5 ) AS right,
"Expands string to length n (2nd argument); trailing blanks produced
"are replaced by the characters from the (3rd) argument
"Note that if n is less than the string, the expression is truncated
"on the right.
"Result: Lufthansa###
rpad( carrname, 12, '#' ) AS rpad,
"All trailing characters that match the character of the 2nd argument
"are removed; trailing blanks are removed, too
"Result: Lufthans
rtrim( carrname, 'a' ) AS rtrim,
"Returns a substring; 2nd argument = position from where to start;
"3rd argument: length of the extracted substring
"Result: fth
substring( carrname, 3, 3 ) AS substring,
"Searches for a PCRE expression and returns the matched substring
"More parameters possible: occurrence, case_sensitive, start, group
"Result: .lu
substring_regexpr( pcre = '\...', "Period that is followed by any two characters
value = url ) AS substring_regexpr,
"All lower case letters are transformed to upper case letters
"Result: LUFTHANSA
upper( carrname ) AS upper
FROM zdemo_abap_carr
WHERE carrid = 'LH'
INTO @DATA(string_functions).Example: Special functions
SELECT SINGLE
carrid,
"Conversion functions
"When used: Special conversions that cannot be handled in a general
"CAST expression
"Type conversion: string of fixed length (e.g. of type c) to variable
"length string of type string
to_clob( carrid ) AS clob,
"Byte string -> character string
bintohex( raw`3599421128650F4EE00008000978B976` ) AS bintohex,
"Character string -> byte string
hextobin( char`3599421128650F4EE00008000978B976` ) AS hextobin,
"Byte field of type RAW to a byte string (BLOB) of type RAWSTRING
to_blob( raw`3599421128650F4EE00008000978B976` ) AS blob,
"Unit and currency conversion functions
"More parameters are available.
"Converts miles to kilometers
unit_conversion( quantity = d34n`1`,
source_unit = unit`MI`,
target_unit = unit`KM` ) AS miles_to_km,
"Converts Euro to US dollars using today's rate
currency_conversion(
amount = d34n`1`,
source_currency = char`EUR`,
target_currency = char`USD`,
exchange_rate_date = @( cl_abap_context_info=>get_system_date( ) )
) AS eur_to_usd,
"Date and time functions
add_days( @( cl_abap_context_info=>get_system_date( ) ), 4 ) AS add_days,
add_months( @( cl_abap_context_info=>get_system_date( ) ), 2 ) AS add_months,
is_valid( @( cl_abap_context_info=>get_system_date( ) ) ) AS date_is_valid,
is_valid( @( cl_abap_context_info=>get_system_time( ) ) ) AS time_is_valid
FROM zdemo_abap_carr
INTO @DATA(special_functions).- Consist of aggregate functions and aggregate the values of multiple rows of the result set of a query into a single value
- See more information here.
Example:
"The example shows a selection of available functions
SELECT
carrid,
"Average value of the content of a column in a row set
AVG( fltime ) AS fltime1,
"AVG with data type specification for the result
AVG( fltime AS DEC( 14,4 ) ) AS fltime2,
"Maximum value of the results in a row set
MAX( fltime ) AS max,
"Minimum value
MIN( fltime ) AS min,
"Sum of the results in a row set.
SUM( fltime ) AS sum,
"Returns the number of rows in a row set.
"The following two have the same meaning.
COUNT( * ) AS count2,
COUNT(*) AS count3,
"Chains the results in a row set.
"An optional separator can be specified
STRING_AGG( airpfrom, ', ' ) AS string_agg
FROM zdemo_abap_flsch
WHERE carrid = 'LH'
GROUP BY carrid
INTO TABLE @DATA(agg_exp).- Arithmetic
expressions
to perform arithmetic calculations using the operators
+,-,*,/, - Cast expressions to convert the value of operands to a dedicated dictionary type. Note that there are special conversion rules.
- String
expressions
using the operator
&&to concatenate character strings. - Case distinctions to carry out either a simple (comparison of the values of a dedicated operand) or complex (searched case; evaluation of multiple logical expressions) case distinction.
The following example demonstrates the expressions mentioned above:
SELECT SINGLE
carrid,
"Arithmethic expressions
"operators + - *
"Note that / is not allowed in integer expressions as the one below
( 1 + 2 ) * 3 AS calc,
"/ used in an expression using type adjustment in ABAP SQL.
"A cast expression converts the value of the operands to the
"specified dictionary type. The result is a representation of the
"source value in the specified type.
CAST( 1 AS D34N ) / CAST( 2 AS D34N ) AS ratio,
"String expression using && to concatenate two character strings;
"the result of the concatenation must not be longer than
"255 characters.
carrid && carrname AS concat,
"Case distinction
"Simple case distinction
"The expression compares the values of an operand with other
"operands. Result: The first operand after THEN for which the
"comparison is true. If no matches are found, the result specified
"after ELSE is selected.
CASE currcode
WHEN 'EUR' THEN 'A'
WHEN 'USD' THEN 'B'
ELSE 'C'
END AS case_simple,
"Complex case distinction
"The expression evaluates logical expressions. Result: The first
"operand after THEN for which the logical expression is true. If no
"logical expressions are true, the result specified after ELSE is
"selected.
CASE WHEN length( carrname ) <= 5 THEN 'small'
WHEN length( carrname ) BETWEEN 6 AND 10 THEN 'mid'
WHEN length( carrname ) BETWEEN 11 AND 15 THEN 'large'
ELSE 'huge'
END AS case_complex
FROM zdemo_abap_carr
WHERE carrid = 'AA'
INTO @DATA(more_sql_expr).How window expressions work:
- Define a subset of the result set (i. e. the "window") of a database query that implements ABAP SQL
- Apply a window
function -
which evaluates the rows of the window and which can, for example,
be an aggregate
function
like
AVGto determine the average value - to the result set - I. e. a window is constructed by the rows of the result set for which all the window functions have the same result; a value is then determined for the rows of a window
Setup of a statement with window expressions:
- Window function, e. g. an aggregate function like
AVG, followed byOVER( ... )(the content in the parentheses defines the "window") - The content in the parentheses can contain the following additions:
- Optional
PARTITION BY: Defines the windows using a comma-separated list of SQL expressions; the window function is calculated for the rows of this window; note that if the addition is not specified, the window comprises all rows of the result set - Optional
ORDER BY: Introduces both an order (you can useASCENDINGandDESCENDING) and a frame (as outlined below) within the current window, which further restricts the rows for which the window function is calculated - A window frame, which stands for a subset of rows inside a
window, can optionally be defined if
ORDER BYis specified; there are 3 options to define the starting and ending frame boundaries (see the example)
- Optional
See more information on window expressions and the syntax here.
Examples:
"Example 1: A simple window is constructed in the OVER clause;
"window functions - here aggregate functions - are applied
SELECT carrid, currency,
SUM( paymentsum ) OVER( PARTITION BY carrid ) AS sum,
AVG( price AS DEC( 14,2 ) ) OVER( PARTITION BY carrid ) AS avg,
MAX( price ) OVER( PARTITION BY carrid ) AS max
FROM zdemo_abap_fli
ORDER BY carrid
INTO TABLE @DATA(win).
"Example 2:
SELECT carrid, currency, fldate,
"Sorts the rows by some columns and counts the number of rows from
"the first row of the window to the current row.
COUNT( * ) OVER( ORDER BY currency, fldate
ROWS BETWEEN
"UNBOUNDED PRECEDING: frame starts at the first row of the window
UNBOUNDED PRECEDING
"CURRENT ROW: determines starting or ending at the current row; here, it ends
AND CURRENT ROW ) AS count1,
"If no window frame is used, the default window frame is
"BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW,
"i. e. the result of count1 equals the result of count2.
COUNT( * ) OVER( ORDER BY currency, fldate ) AS count2,
"Sorts the rows by some columns and counts the number of rows from
"the current row to the last row of the window.
"The result is reverse numbering.
COUNT( * ) OVER( ORDER BY currency, fldate
ROWS BETWEEN CURRENT ROW
UNBOUND FOLLOWING:
"Determines the ending frame boundary, this addition specifies the last row of the window
AND UNBOUNDED FOLLOWING ) AS count_reverse,
"Sorts the rows by some columns and calculates the rolling averages
"of a subset of rows from column price. The subset consists of the
"current row plus one preceding and one following row. Another use
"case as below example that uses prices would be that, for example,
"you can calculate the 3-day-average temperature for every day from
"a list of temperature data.
AVG( price AS DEC( 14,2 ) ) OVER( ORDER BY currency, fldate
ROWS BETWEEN
"n PRECEDING: for both start and end of frame; frame to start/end n rows above the current row
1 PRECEDING
"n FOLLOWING: for both start and end of frame; frame to start/end n rows beneath the current row
AND 1 FOLLOWING ) AS avg
FROM zdemo_abap_fli
INTO TABLE @DATA(result).You can formulate conditions in ABAP SQL statements, i. e. logical
expressions,
especially in the WHERE clause to restrict the result. Note
that without a WHERE clause, all rows are respected for the
operation.
See below a selection of the operators that are possible when specifying conditions. For more information, see the subtopics of the SQL Conditions topic.
| Operator | Meaning |
|---|---|
=, EQ |
The content of two operands is equal. |
<>, NE |
The content of two operands is not equal. |
<, LT |
The content of one operand is less than the content of the other operand. |
>, GT |
The content of one operand is greater than the content of the other operand. |
<=, LE |
The content of one operand is less than or equal to the content of the other operand. |
>=, GE |
The content of one operand is greater than or equal to the content of the other operand. |
... [NOT] BETWEEN ... AND ... |
The value of an operand is (not) between the value of the two other operands. |
... [NOT] LIKE ... |
The content of an operand matches (does not match) a specified pattern. The pattern can be specified by using wildcard characters. % stands for any character string, including an empty string. _ stands for any character. |
... IS [NOT] INITIAL ... |
The value of an operand is (not) the initial value of its built-in dictionary type. |
... EXISTS ... |
Checks the result set of a subquery. The expression is true if the result set contains at least one row. See more information here. |
... [NOT] IN ... |
Checks whether the operands on the left side match a value from a set of values specified in parentheses. On the left side, a single operand or an operand list are possible. On the right side, a comma-separated lists or subqueries can be specified. It is also possible to specify a ranges table to evaluate ranges conditions. |
... IS [NOT] NULL ... |
Checks whether the value of an operand is (not) the null value. Find more information in the code snippet and in the ABAP Keyword Documentation. |
💡 Note
You can combine multiple logical expressions into one logical expression usingANDorOR. To further detail out the desired condition, expressions within parentheses are possible.
Examples:
"---- SQL conditions demonstrated with the WHERE clause ----
"Note:
"- For most of the self-contained examples, an internal table is used as the
" data source of SELECT statements to work with simple data.
"- For some examples that are covered, such as subqueries, demo database tables
" from the cheat sheet repository are used in addition.
"- Dynamic specifications are also possible. They are not covered here. See
" the Dynamic Programming cheat sheet.
"---- Types and internal table to work with in the examples ----
"Note: You cannot use type string columns in WHERE conditions.
TYPES: BEGIN OF demo_struc,
id TYPE i,
name TYPE c LENGTH 15,
"name TYPE string,
END OF demo_struc.
DATA itab TYPE SORTED TABLE OF demo_struc WITH UNIQUE KEY id.
"Populating internal table with data to work with in the examples
itab = VALUE #( ( id = 1 name = 'bear' )
( id = 2 name = 'camel' )
( id = 3 name = 'rabbit' )
( id = 4 name = 'zebra' )
( id = 5 name = 'dog' )
( id = 6 name = 'deer' )
( id = 7 name = 'squirrel' )
( id = 8 name = 'cheetah' )
( id = 9 name = 'elephant' )
( id = 10 name = 'donkey' )
( id = 11 name = 'fish' )
( id = 12 name = 'sheep' ) ).
"---- =, <>, >, >= (as a selection of possible comparison operators) ----
SELECT id FROM @itab AS tab WHERE name = 'bear' INTO TABLE @DATA(it). "1
SELECT id FROM @itab AS tab WHERE name <> 'bear' INTO TABLE @it. "2-12
SELECT id FROM @itab AS tab WHERE id > 10 INTO TABLE @it. "11,12
SELECT id FROM @itab AS tab WHERE id >= 10 INTO TABLE @it. "10,11,12
"---- Combining logical expressions using AND, OR and parentheses ----
SELECT id FROM @itab AS tab WHERE id = 1 AND name = 'bear' INTO TABLE @it. "1
SELECT id FROM @itab AS tab WHERE name = 'bear' OR name = 'sheep' INTO TABLE @it. "1,12
"In the following example, the resulting table is initial. One of the expressions
"in parentheses is false (AND is used between the expressions in parentheses).
"In contrast, the example below returns an entry because of using OR.
SELECT id FROM @itab AS tab
WHERE ( id = 1 AND name = 'bear' )
AND ( id = 20 AND name = 'camel' )
INTO TABLE @it.
SELECT id FROM @itab AS tab
WHERE ( id = 1 AND name = 'bear' )
OR ( id = 20 AND name = 'camel' )
INTO TABLE @it. "1
"------------------------ [NOT] BETWEEN ------------------------
SELECT id FROM @itab AS tab WHERE id BETWEEN 1 AND 4 INTO TABLE @it. "1,2,3,4
"The condition with BETWEEN above corresponds to the following condition.
"The example makes use of a condition specified in parentheses to combine multiple
"expressions.
SELECT id FROM @itab AS tab WHERE ( id >= 1 AND id <= 4 ) INTO TABLE @it. "1,2,3,4
"Negation with NOT
SELECT id FROM @itab AS tab WHERE id NOT BETWEEN 1 AND 4 INTO TABLE @it. "5-12
"------------------------ IS [NOT] INITIAL ------------------------
SELECT id FROM @itab AS tab WHERE id IS NOT INITIAL INTO TABLE @it. "1-12
"------------------------ [NOT] LIKE ------------------------
"For (not) matching a specified pattern
"Note: % (any character string), _ (any character).
SELECT name FROM @itab AS tab
WHERE name LIKE '%ee%'
OR name LIKE '_o%'
INTO TABLE @DATA(names). "dog,deer,cheetah,donkey,sheep
"ESCAPE addition for defining a single-character escape character
"In the following example, this character is #. It is placed before
"the % character in the specification after LIKE. In this case, %
"is escaped and does then not stand for any character string in the
"evaluation.
"Adding a table entry for this syntax example.
itab = VALUE #( BASE itab ( id = 13 name = '100%' ) ).
"Any character sequence followed by the % character
SELECT name FROM @itab AS tab
WHERE name LIKE '%#%' ESCAPE '#'
INTO TABLE @names. "100%
"Deleting the entry because it is not relevant for the further examples.
DELETE itab INDEX 13.
"------------------------ [NOT] IN (using a value set) ------------------------
"For (not) matching a value in a set of values specified in parentheses.
"Single operands on the left side of IN
SELECT id FROM @itab AS tab
WHERE name IN ( 'camel', 'rabbit', 'dog', 'snake' )
INTO TABLE @it. "2,3,5
"Negation NOT IN; note to use host variables/expressions for local/global data objects
DATA(animal) = 'sheep'.
SELECT id FROM @itab AS tab
WHERE name NOT IN ( 'fish', @animal )
INTO TABLE @it. "1-10
"Operand list (a parenthesized comma-separated list) on the left side of IN
"For (not) matching value tuples from a set of value tuples specified in parentheses on the right side.
"In the following example, two values are specified in the operand list on the left. Consequently,
"two values with appropriate types must be specified in parentheses on the right.
SELECT id FROM @itab AS tab
WHERE ( id, name ) IN ( ( 1, 'bear' ), ( 3, 'rabbit' ), ( 8, 'zebra' ), ( 20, 'dog' ) )
INTO TABLE @it. "1,3
"------------------------ [NOT] IN (using a subquery) ------------------------
"[NOT] IN for matching a value contained in the result set of a subquery
"In the following example, the subquery reads data from a demo database table.
"For a representative result, the table is cleared, and then filled with 'suitable'
"data sets.
DELETE FROM zdemo_abap_tab1.
MODIFY zdemo_abap_tab1 FROM TABLE @( VALUE #( ( key_field = 11 num1 = 11 )
( key_field = 12 num1 = 12 )
( key_field = 13 num1 = 13 )
( key_field = 14 num1 = 14 ) ) ).
SELECT id FROM @itab AS tab
WHERE id IN ( SELECT key_field FROM zdemo_abap_tab1 ) INTO TABLE @it. "11,12
"------------------------ [NOT] IN (using a ranges table) ------------------------
"[NOT] IN for checking whether the operands on the left side match a ranges condition in a ranges table
"Declaring a ranges table
DATA rangestab TYPE RANGE OF i.
"Populating a ranges table using the VALUE operator
rangestab = VALUE #( ( sign = 'I' option = 'BT' low = 1 high = 3 )
( sign = 'I' option = 'GE' low = 10 ) ).
SELECT id FROM @itab AS tab WHERE id IN @rangestab INTO TABLE @it. "1,2,3,10,11,12
"You cannot use logical operators such as CP (conforms to pattern) in the WHERE clause.
"In a ranges table, they are possible.
"Note:
"- Regarding CP: * (any character sequence), + (any character), # (escape character)
"- An equivalent example above uses the LIKE addition.
DATA rt TYPE RANGE OF demo_struc-name.
rt = VALUE #( ( sign = 'I' option = 'CP' low = '*ee*' ) "ee in a string
( sign = 'I' option = 'CP' low = '+o*' ) ). "o in second position
SELECT name FROM @itab AS tab
WHERE name IN @rt
INTO TABLE @names. "dog,deer,cheetah,donkey,sheep
"------------------------ EXISTS ------------------------
"For checking the result set of a subquery.
"The following example reads all entries from the internal table if entries having
"the same key also exist in the database table.
"Note: The SELECT list in the subquery only contains a literal to determine that
"the entry exists. Specifying explicit column names is not relevant.
SELECT id FROM @itab AS tab WHERE
EXISTS ( SELECT @abap_true FROM zdemo_abap_tab1 WHERE key_field = tab~id )
INTO TABLE @it. "11,12
"------------------------ IS [NOT] NULL ------------------------
"The null value is a special value that is returned by a database. It indicates an
"undefined value or result. Note that, in ABAP, there are no special null values. Do
"not confuse the null value with a type-dependent initial value. When using SELECT
"statements to read data, null values can be produced by, for example, outer joins.
"When the null values are passed to a data object, they are transformed to the
"type-dependent initial values. For more information, refer to the ABAP Keyword Documentation.
"The following example uses a left outer join to intentionally create null values. For
"this purpose, two demo database tables of the cheat sheet repository are cleared and
"populated with specific values to visualize null values.
DELETE FROM zdemo_abap_tab1.
DELETE FROM zdemo_abap_tab2.
MODIFY zdemo_abap_tab1 FROM TABLE @( VALUE #( ( key_field = 1 char1 = 'a' char2 = 'y' )
( key_field = 2 char1 = 'b' char2 = 'z' ) ) ).
MODIFY zdemo_abap_tab2 FROM TABLE @( VALUE #( ( key_field = 1 char1 = 'a' )
( key_field = 2 char1 = 'a' )
( key_field = 3 char1 = 'b' )
( key_field = 4 ) ) ).
"Note that for the entry 'key_field = 4' no char1 value was passed.
"char1 is a shared column of the two database tables, and which is used in
"the ON condition of the join. Since there is no entry in char1 for 'key_field = 4',
"the joined values are null in that case. The WHERE clause uses the addition IS NULL.
"Therefore, the result only contains this entry. char2 is assigned the type-initial
"value in the result.
SELECT tab2~key_field, tab1~char2
FROM zdemo_abap_tab2 AS tab2
LEFT OUTER JOIN zdemo_abap_tab1 AS tab1 ON tab1~char1 = tab2~char1
WHERE tab1~char1 IS NULL
INTO TABLE @DATA(joined_tab).
*KEY_FIELD CHAR2
*4
"The following example visualizes the null values. The INDICATORS addition of the
"INTO clause is used to specify indicators such as the null indicator. In the
"example, an appropriate target table is defined to also store information about
"which columns of the result set contain the null value and which do not.
"For more information on the syntax, refer to the ABAP Keyword Documentation.
TYPES: BEGIN OF st4null,
BEGIN OF s2,
key_field TYPE zdemo_abap_tab2-key_field,
char2 TYPE zdemo_abap_tab1-char2,
END OF s2,
BEGIN OF nulls,
key_field TYPE c LENGTH 1,
char2 TYPE c LENGTH 1,
END OF nulls,
END OF st4null.
DATA joined_tab_w_null_ind TYPE TABLE OF st4null WITH EMPTY KEY.
SELECT tab2~key_field, tab1~char2
FROM zdemo_abap_tab2 AS tab2
LEFT OUTER JOIN zdemo_abap_tab1 AS tab1 ON tab1~char1 = tab2~char1
INTO TABLE @joined_tab_w_null_ind INDICATORS NULL STRUCTURE nulls.
*S2 NULLS
*KEY_FIELD CHAR2 KEY_FIELD CHAR2
*1 y
*KEY_FIELD CHAR2 KEY_FIELD CHAR2
*2 y
*KEY_FIELD CHAR2 KEY_FIELD CHAR2
*3 z
*KEY_FIELD CHAR2 KEY_FIELD CHAR2
*4 X
"Negation IS NOT NULL
SELECT tab2~key_field, tab1~char2
FROM zdemo_abap_tab2 AS tab2
LEFT OUTER JOIN zdemo_abap_tab1 AS tab1 ON tab1~char1 = tab2~char1
WHERE tab1~char1 IS NOT NULL
INTO TABLE @joined_tab.
*KEY_FIELD CHAR2
*1 y
*2 y
*3 zFOR ALL ENTRIES
addition:
- Components of an internal table can be used in the
WHEREclause in logical expressions for comparisons with a column of the data source. - The logical expression is evaluated for each individual row of the internal table.
- The result set of the
SELECTstatement is the union set of the result sets produced by the individual evaluations. Rows that occur more than once are removed from the result set automatically. The entire content of a row is respected. - If
FOR ALL ENTRIESis specified, there must be at least one comparison with a column of the internal table. - For more information, especially restricitions and things to pay attention to (e. g. making sure that the internal table is not initial), see here.
"Checking that table is not initial
IF ( 0 < lines( itab2 ) ).
SELECT comp1, comp2, comp3
FROM dbtab
FOR ALL ENTRIES IN @itab2 "Host variable before internal table
WHERE comp1 = @itab2-comp1 ... "Relational expression on the right side of a comparison
INTO TABLE @itab1
ENDIF.Checking the result set of a subquery with the addition EXISTS
See possible clauses and additions of a subquery in a condition in ABAP SQL here.
The following code snippet includes a parenthesized subquery following EXISTS. Data is only selected from dbtab1 if the relational expression (WHERE EXISTS ...) is true, i. e. if the result set of the subquery contains at least one row. Note the components of the table that are referenced using a tilde.
SELECT comp1, comp2, comp3
FROM dbtab1 AS tab1
WHERE EXISTS
( SELECT comp1 FROM dbtab2
WHERE comp1 = tab1~comp1 AND comp2 = tab1~comp2 )
INTO ...Using an inner join:
- Columns of two or more data sources in a result set can be joined.
- Result set:
- Columns of the rows in the result set of the left side with the columns of the rows in the result set of the right side are joined into a single result set.
- Contains all combinations of rows for whose columns the join condition is true.
- If there are identical column names in multiple data sources, use the column
selector
~.
SELECT a~comp1, a~comp2, b~comp3, c~comp4
FROM dbtab1 AS a
INNER JOIN dbtab2 AS b ON a~comp1 = b~comp1 AND a~comp2 = b~comp2
INNER JOIN dbtab3 AS c ON a~comp1 = c~comp1
WHERE ...
INTO ...Using an outer join:
- Realized by either a left outer join or a right outer join.
- Result set:
- Same result set as the inner join.
- Difference: For each selected row on the left side as
LEFT OUTER JOINor on the right side asRIGHT OUTER JOIN, at least one row is created in the result set even if no rows on the other side meet the condition. The columns on the other side that do not meet the condition are filled with null values.
"Example for a left outer join
SELECT a~comp1, a~comp2, b~comp3,
FROM dbtab1 AS a
LEFT OUTER JOIN dbtab2 AS b ON a~comp1 = b~comp1
WHERE ...
INTO ...💡 Note
There are more join variants available. See the ABAP Keyword Documentation on joins for more information.
Merging the result sets of multiple queries into a single result set using the set operator UNION. In this case, the rows of the result set of the query after UNION are inserted into the result set of the query in front of UNION.
SELECT FROM dbtab1
FIELDS ...
WHERE ...
UNION
SELECT FROM dbtab2
FIELDS ...
WHERE ...
INTO ...When to use Common Table Expressions (CTE):
- Whenever you need intermediate results in a
SELECTstatement and especially if you need them more than once. - You get the option of selecting directly from a subquery (
SELECT FROM subquery), which is not possible in ABAP SQL.
How it works:
- The ABAP SQL keyword
WITHintroduces the definition of CTEs. - Each CTE creates a tabular result set in a subquery.
- The result set of such a CTE can then be used in subsequent queries as data source; CTEs can be considered as temporary views, which only exist for the duration of the database access.
- The CTEs (at least one) are then used in a final main
query, i.
e. a
SELECTstatement accesses the result of the expressions.
Setup of a statement with CTE:
- Introductory keyword
WITH - A comma-separated list with at least one definition of a CTE
- Each CTE has a unique name with an initial
+character - An optional list of column names, which should be used in the result set, within parentheses
ASfollowed by a subquery withSELECTwhich creates the tabular result set of the CTE
- Each CTE has a unique name with an initial
- A closing main query with
SELECTin which the previous CTEs are to be used as data source - If a
SELECTloop is opened and data is written into a work area in the closing main query, the loop must be closed withENDWITH.(which fulfills the same task asENDSELECT.).
💡 Note
- Each CTE must be used at least once, either in another CTE or in the main query. The main query must access at least one CTE.
- The result set of a CTE never has a client column.
- See more information in this topic and further options and additions when using CTEs in the subtopics.
Example: The result sets of both common table expressions
+connections and +sum_seats are merged in the
subquery of the CTE +result in a join expression. An explicit
name list assigns names to the resulting columns. These names are used
in the main query to sort the results. For each flight connection of the
selected airline, the total number of occupied seats is stored in the
internal table.
WITH
+connections AS (
SELECT zdemo_abap_flsch~carrid, carrname, connid, cityfrom, cityto
FROM zdemo_abap_flsch
INNER JOIN zdemo_abap_carr
ON zdemo_abap_carr~carrid = zdemo_abap_flsch~carrid
WHERE zdemo_abap_flsch~carrid BETWEEN 'AA' AND 'JL' ),
+sum_seats AS (
SELECT carrid, connid, SUM( seatsocc ) AS sum_seats
FROM zdemo_abap_fli
WHERE carrid BETWEEN 'AA' AND 'JL'
GROUP BY carrid, connid ),
+result( name, connection, departure, arrival, occupied ) AS (
SELECT carrname, c~connid, cityfrom, cityto, sum_seats
FROM +connections AS c
INNER JOIN +sum_seats AS s
ON c~carrid = s~carrid AND c~connid = s~connid )
SELECT *
FROM +result
ORDER BY name, connection
INTO TABLE @DATA(result).- Inserts one or more rows into a database table specified.
- The rows to be inserted are taken from a structure, an internal table, or the result set of an embedded subquery.
- As mentioned above, structures and internal tables from which to insert content should be specified as host variables (with
@) or host expressions (with@( ... )). - The system fields
sy-subrc(0 = single row or all rows inserted successfully, 4 = row not or not all rows inserted) andsy-dbcnt(number of rows that are inserted) are set. - More information:
INSERT
"Inserting a single row into a database table
INSERT dbtab FROM @row.
"Alternative syntax, same effect
INSERT INTO dbtab VALUES @row.
"Line is created inline using the VALUE operator as part of a host expression
INSERT dbtab FROM @( VALUE #( comp1 = ... comp2 = ... ) ).
"Inserting multiple lines from an internal table into a database table.
"Make sure that the internal table does not contain a line having the same key
"as an existing row in the database table. Otherwise, a runtime error occurs.
INSERT dbtab FROM TABLE @itab.
"Inserting lines from a table declared inline using the VALUE operator
"as part of a host expression
INSERT dbtab FROM TABLE @( VALUE #( ( comp1 = ... comp2 = ... )
( comp1 = ... comp2 = ... ) ) ).
"ACCEPTING DUPLICATE KEYS addition: To avoid the runtime error mentioned above,
"all lines that would produce duplicate entries in the database table
"regarding the keys are discarded and sy-subrc is set to 4.
INSERT dbtab FROM TABLE @itab ACCEPTING DUPLICATE KEYS.
"Inserting the result set of an embedded subquery
"Here, multiple result sets can be joined, e. g. using UNION.
INSERT dbtab FROM ( SELECT ... ).- Changes the content of one or more rows of a database table specified.
- Similar to
INSERT,sy-subrcandsy-dbcntare set. - More information:
UPDATE
"Changing content by overwriting entire rows based on a structure
UPDATE dbtab FROM @row.
UPDATE dbtab FROM @( VALUE #( comp1 = ... comp2 = ... ) ). "Using a host expression
"Changing content by overwriting entire rows based on rows in an internal table
UPDATE dbtab FROM TABLE @itab.
"Using a host expression
UPDATE dbtab FROM TABLE @( VALUE #( ( comp1 = ... comp2 = ... )
( comp1 = ... comp2 = ... ) ) ).
"INDICATORS addition: Changing content of specific fields without overwriting
"existing values of other fields
"Example:
"- Structured type is created with WITH INDICATORS addition
"- Internal table from which to update dbtab is created;
" it includes the indicator structure comp_ind
"- Internal table is filled; only one component is flagged as to be updated
"- Other fields remain unchanged; note that key fields must be included
" in ind_tab (indicator setting for key fields has no effect)
TYPES ind_wa TYPE dbtab WITH INDICATORS comp_ind TYPE abap_bool.
DATA ind_tab TYPE TABLE OF ind_wa.
ind_tab = VALUE #(
( comp1 = ... comp2 = ... comp_ind-comp2 = abap_true )
( comp1 = ... comp2 = ... comp_ind-comp2 = abap_true ) ).
UPDATE dbtab FROM TABLE @ind_tab INDICATORS SET STRUCTURE comp_ind.
"In the following example, the logic is reversed using NOT.
UPDATE dbtab FROM TABLE @ind_tab INDICATORS NOT SET STRUCTURE comp_ind.
"SET addition: Changing values of specific fields in all table rows
"There are mutliple options for the value assignment. E.g. you can use
"a literal, host variable/expression, SQL function, and so on.
UPDATE dbtab SET comp2 = ... .- Inserts one or more rows into a database table specified or overwrites existing ones.
- As above,
sy-subrcandsy-dbcntare set. - More information:
MODIFY
"Inserting a single row into a database table or changing an existing row
MODIFY dbtab FROM @row.
"Using a host expression
MODIFY dbtab FROM @( VALUE #( comp1 = ... comp2 = ... ) ).
"Inserting/Changing multiple rows
MODIFY dbtab FROM TABLE @itab.
"Using a host expression
MODIFY dbtab FROM TABLE @( VALUE #( ( comp1 = ... comp2 = ... )
( comp1 = ... comp2 = ... ) ) ).
"Inserting/Changing multiple rows based on a result set of an embedded subquery
MODIFY dbtab FROM ( SELECT ... ).- Deletes one or more rows from a database table specified.
- As above,
sy-subrcandsy-dbcntare set. - More information:
DELETE
"Variant DELETE FROM ...: Either all rows are deleted or restricted
"All rows are deleted
DELETE FROM dbtab.
"Rows are deleted based on a condition
DELETE FROM dbtab WHERE ....
"Note that there are further options available, e. g. ORDER BY, UP TO
"Variant DELETE ... FROM ...: Deleting a single row or multiple row
DELETE dbtab FROM @row.
"Using a host expression
DELETE dbtab FROM @( VALUE #( comp1 = ... ) ).
DELETE dbtab FROM TABLE @itab.
"Using a host expression
DELETE dbtab FROM TABLE @( VALUE #( ( comp1 = ... )
( comp1 = ... ) ) ).Find more information in the ABAP for RAP: Entity Manipulation Language (ABAP EML) cheat sheet.
- Note that ABAP SQL statements offer syntax options for dynamic programming. For example, you can specify the data source to read from dynamically. See more information in the ABAP Keyword Documentation or the ABAP cheat sheet on dynamic programming.
DATA(dbtab) = 'ZDEMO_ABAP_FLSCH'. "Selecting from a dynamically specified database table. SELECT * FROM (dbtab) WHERE ... INTO ...
- This topic serves as the entry point for topics about ABAP SQL in the ABAP Keyword Documentation. For the full details, check the subtopics there, especially topics not covered in this cheat sheet.
- There are RAP-specific variants of ABAP SQL statements that use the
MAPPING FROM ENTITYaddition. Find more information here.
💡 Note
- The executable example covers the following topics, among others:
- Reading from database tables using
SELECT- Various additions to
SELECTstatements- Changing data in database tables using
INSERT,UPDATE,MODIFYandDELETE- Excursions: Operands and expressions in ABAP SQL statements
- The steps to import and run the code are outlined here.
- Disclaimer