Related
I’m fairly used to using MySQL, but not particularly familiar with SQL Server. Tough luck, the database I’m dealing with here is on SQL Server 2014.
I have a table with a column whose values are all integers with leading, separating, and trailing semicolons, like these three fictitious rows:
;905;1493;384;13387;29;933;467;28732;
;905;138;3084;1387;290;9353;4767;2732;
;9085;14493;3864;130387;289;933;4767;28732;
What I am trying to do now is to select all rows where more than one number taken from a list of numbers appears in this column. So for example, given the three rows above, if I have the group 905,467,4767, the statement I’m trying to figure out how to construct should return the first two rows: the first row contains 905 and 467; the second row contains 905 and 4767. The third row contains only 4767, so that row should not be returned.
As far as I can tell, SQL Server does not actually support regex directly (and I don’t even know what managed code is), which doesn’t help. Even with regex, I wouldn’t know where to begin. Oracle seems to have a function that would be very useful, but that’s Oracle.
Most similar questions on here deal with finding multiple instances of the same character (usually singular) and solve the problem by replacing the string to match with nothing and counting the difference in length. I suppose that would technically work here, too, but given a ‘filter’ group of 15 numbers, the SELECT statement would become ridiculously long and convoluted and utterly unreadable. Additionally, I only want to match entire numbers (so if one of the numbers to match is 29, the value 29 would match in the first row, but the value 290 in the second row should not match), which means I’d have to include the semicolons in the REPLACE clause and then discount them when calculating the length. A complete mess.
What I would ideally like to do is something like this:
SELECT * FROM table WHERE REGEXP_COUNT(column, ';(905|467|4767);') > 1
– but that will obviously not work, for all kinds of reasons (the most obvious one being the nonexistence of REGEXP_COUNT outside Oracle).
Is there some sane, manageable way of doing this?
You can do
SELECT *
FROM Mess
CROSS APPLY (SELECT COUNT(*)
FROM (VALUES (905),
(467),
(4767)) V(Num)
WHERE Col LIKE CONCAT('%;', Num, ';%')) ca(count)
WHERE count > 1
SQL Fiddle
Or alternatively
WITH Nums
AS (SELECT Num
FROM (VALUES (905),
(467),
(4767)) V(Num))
SELECT Mess.*
FROM Mess
CROSS APPLY (VALUES(CAST(CONCAT('<x>', REPLACE(Col, ';', '</x><x>'), '</x>') AS XML))) x(x)
CROSS APPLY (SELECT COUNT(*)
FROM (SELECT n.value('.', 'int')
FROM x.x.nodes('/x') n(n)
WHERE n.value('.', 'varchar') <> ''
INTERSECT
SELECT Num
FROM Nums) T(count)
HAVING COUNT(*) > 1) ca2(count)
Could you put your arguments into a table (perhaps using a table-valued function accepting a string (of comma-separated integers) as a parameter) and use something like this?
DECLARE #T table (String varchar(255))
INSERT INTO #T
VALUES
(';905;1493;384;13387;29;933;467;28732;')
, (';905;138;3084;1387;290;9353;4767;2732;')
, (';9085;14493;3864;130387;289;933;4767;28732;')
DECLARE #Arguments table (Arg int)
INSERT INTO #Arguments
VALUES
(905)
, (467)
, (4767)
SELECT String
FROM
#T
CROSS JOIN #Arguments
GROUP BY String
HAVING SUM(CASE WHEN PATINDEX('%;' + CAST(Arg AS varchar) + ';%', String) > 0 THEN 1 ELSE 0 END) > 1
And example of using this with a function to generate the arguments:
CREATE FUNCTION GenerateArguments (#Integers varchar(255))
RETURNS #Arguments table (Arg int)
AS
BEGIN
WITH cte
AS
(
SELECT
PATINDEX('%,%', #Integers) p
, LEFT(#Integers, PATINDEX('%,%', #Integers) - 1) n
UNION ALL
SELECT
CASE WHEN PATINDEX('%,%', SUBSTRING(#Integers, p + 1, LEN(#Integers))) + p = p THEN 0 ELSE PATINDEX('%,%', SUBSTRING(#Integers, p + 1, LEN(#Integers))) + p END
, CASE WHEN PATINDEX('%,%', SUBSTRING(#Integers, p + 1, LEN(#Integers))) = 0 THEN RIGHT(#Integers, PATINDEX('%,%', REVERSE(#Integers)) - 1) ELSE LEFT(SUBSTRING(#Integers, p + 1, LEN(#Integers)), PATINDEX('%,%', SUBSTRING(#Integers, p + 1, LEN(#Integers))) - 1) END
FROM cte
WHERE p <> 0
)
INSERT INTO #Arguments (Arg)
SELECT n
FROM cte
RETURN
END
GO
DECLARE #T table (String varchar(255))
INSERT INTO #T
VALUES
(';905;1493;384;13387;29;933;467;28732;')
, (';905;138;3084;1387;290;9353;4767;2732;')
, (';9085;14493;3864;130387;289;933;4767;28732;')
;
SELECT String
FROM
#T
CROSS JOIN GenerateArguments('905,467,4767')
GROUP BY String
HAVING SUM(CASE WHEN PATINDEX('%;' + CAST(Arg AS varchar) + ';%', String) > 0 THEN 1 ELSE 0 END) > 1
You can achieve this using the like function for the regex and row_number to determine the number of matches.
Here we declare the column values for testing:
DECLARE #tbl TABLE (
string NVARCHAR(MAX)
)
INSERT #tbl VALUES
(';905;1493;384;13387;29;933;467;28732;'),
(';905;138;3084;1387;290;9353;4767;2732;'),
(';9085;14493;3864;130387;289;933;4767;28732;')
Then we pass your search parameters into a table variable to be joined on:
DECLARE #search_tbl TABLE (
search_value INT
)
INSERT #search_tbl VALUES
(905),
(467),
(4767)
Finally we join the table with the column to search for onto the search table. We apply the row_number function to determine the number of times it matches. We select from this subquery where the row_number = 2 meaning that it joined at least twice.
SELECT
string
FROM (
SELECT
tbl.string,
ROW_NUMBER() OVER (PARTITION BY tbl.string ORDER BY tbl.string) AS rn
FROM #tbl tbl
JOIN #search_tbl search_tbl ON
tbl.string LIKE '%;' + CAST(search_tbl.search_value AS NVARCHAR(MAX)) + ';%'
) tbl
WHERE rn = 2
You could build a where clause like this :
WHERE
case when column like '%;905;%' then 1 else 0 end +
case when column like '%;467;%' then 1 else 0 end +
case when column like '%;4767;%' then 1 else 0 end >= 2
The advantage is that you do not need a helper table. I don't know how you build the query, but the following also works, and is useful if the numbers are in a tsql variable.
case when column like ('%;' + #n + ';%') then 1 else 0 end
I'm converting a stored procedure from MySql to SQL Server. The procedure has one input parameter nvarchar/varchar which is a comma-separated string, e.g.
'1,2,5,456,454,343,3464'
I need to write a query that will retrieve the relevant rows, in MySql I'm using FIND_IN_SET and I wonder what the equivalent is in SQL Server.
I also need to order the ids as in the string.
The original query is:
SELECT *
FROM table_name t
WHERE FIND_IN_SET(id,p_ids)
ORDER BY FIND_IN_SET(id,p_ids);
The equivalent is like for the where and then charindex() for the order by:
select *
from table_name t
where ','+p_ids+',' like '%,'+cast(id as varchar(255))+',%'
order by charindex(',' + cast(id as varchar(255)) + ',', ',' + p_ids + ',');
Well, you could use charindex() for both, but the like will work in most databases.
Note that I've added delimiters to the beginning and end of the string, so 464 will not accidentally match 3464.
You would need to write a FIND_IN_SET function as it does not exist. The closet mechanism I can think of to convert a delimited string into a joinable object would be a to create a table-valued function and use the result in a standard in statement. It would need to be similar to:
DECLARE #MyParam NVARCHAR(3000)
SET #MyParam='1,2,5,456,454,343,3464'
SELECT
*
FROM
MyTable
WHERE
MyTableID IN (SELECT ID FROM dbo.MySplitDelimitedString(#MyParam,','))
And you would need to create a MySplitDelimitedString type table-valued function that would split a string and return a TABLE (ID INT) object.
A set based solution that splits the id's into ints and join with the base table which will make use of index on the base table id. I assumed the id would be an int, otherwise just remove the cast.
declare #ids nvarchar(100) = N'1,2,5,456,454,343,3464';
with nums as ( -- Generate numbers
select top (len(#ids)) row_number() over (order by (select 0)) n
from sys.messages
)
, pos1 as ( -- Get comma positions
select c.ci
from nums n
cross apply (select charindex(',', #ids, n.n) as ci) c
group by c.ci
)
, pos2 as ( -- Distinct posistions plus start and end
select ci
from pos1
union select 0
union select len(#ids) + 1
)
, pos3 as ( -- add row number for join
select ci, row_number() over (order by ci) as r
from pos2
)
, ids as ( -- id's and row id for ordering
select cast(substring(#ids, p1.ci + 1, p2.ci - p1.ci - 1) as int) id, row_number() over (order by p1.ci) r
from pos3 p1
inner join pos3 p2 on p2.r = p1.r + 1
)
select *
from ids i
inner join table_name t on t.id = i.id
order by i.r;
You can also try this by using regex to get the input values from comma separated string :
select * from table_name where id in (
select regexp_substr(p_ids,'[^,]+', 1, level) from dual
connect by regexp_substr(p_ids, '[^,]+', 1, level) is not null );
I have a table with values in a field called 'code'.
ABC
DFG
CDF
How would I select all codes that are not in the table from a list I have?
Eg:
SELECT * from [my list] where table1.code not in [my list]
the list is not in a table.
The list would be something like "ABC","BBB","TTT" (As strings)
Try this:
SELECT code
FROM Table1
WHERE code NOT IN ('ABC','CCC','DEF') --values from your list
It will result:
DFG
CDF
If the list is in another table, try this:
SELECT code
FROM Table1
WHERE code NOT IN (SELECT code FROM Table2)
As per your requirement, try this:
SELECT list
FROM Table2
WHERE list NOT IN (SELECT code from table1)
It will select the list values that are not in code.
See an example in SQL Fiddle
The question key point need to set "ABC","BBB","TTT" source data trun to a table.
that table will look like
|---+
|val|
|---+
|ABC|
|BBB|
|TTT|
Sqlite didn't support sqlite function. so that will be a little hard to sqlite your list to be a table.
You can use a CTE Recursive to make like sqlite function
You need to use replace function to remove " double quotes from your
source data.
There are two column in the CTE
val column carry your List data
rest column to remember current splite string
You will get a table from CTE like this.
|---+
|val|
|---+
|ABC|
|BBB|
|TTT|
Then you can compare the data with table1.
Not IN
WITH RECURSIVE split(val, rest) AS (
SELECT '', replace('"ABC","BBB","TTT"','"','') || ','
UNION ALL
SELECT
substr(rest, 0, instr(rest, ',')),
substr(rest, instr(rest, ',')+1)
FROM split
WHERE rest <> '')
SELECT * from (
SELECT val
FROM split
WHERE val <> ''
) t where t.val not IN (
select t1.code
from table1 t1
)
sqlfiddle:https://sqliteonline.com/#fiddle-5adeba5dfcc2fks5jgd7ernq
Outut Result:
+---+
|val|
+---+
|BBB|
|TTT|
If you want to show it in a line,use GROUP_CONCAT function.
WITH RECURSIVE split(val, rest) AS (
SELECT '', replace('"ABC","BBB","TTT"','"','') || ','
UNION ALL
SELECT
substr(rest, 0, instr(rest, ',')),
substr(rest, instr(rest, ',')+1)
FROM split
WHERE rest <> '')
SELECT GROUP_CONCAT(val,',') val from (
SELECT val
FROM split
WHERE val <> ''
) t where t.val not IN (
select t1.code
from table1 t1
)
Outut Result:
BBB, TTT
sqlfiddle:https://sqliteonline.com/#fiddle-5adecb92fcc36ks5jgda15yq
Note:That is unreasonable on SELECT * from [my list] where table1.code not in [my list],because This query has no place to find table1 so you couldn't get table1.code column
You can use not exists or JOIN to make your expect.
sqlfiddle:https://sqliteonline.com/#fiddle-5adeba5dfcc2fks5jgd7ernq
Can you use common table expressions?
WITH temp(code) AS (VALUES('ABC'),('BBB'),('TTT'),(ETC...))
SELECT temp.code FROM temp WHERE temp.code NOT IN
(SELECT DISTINCT table1.code FROM table1);
This would allow you to create a temporary table defined with your list of strings within the VALUES statement. Then use standard SQL to select values NOT IN your table1.code column.
Is this solution good, or am I missing something?
create table table10 (code varchar(20));
insert into table10 (code) values ('ABC');
insert into table10 (code) values ('DFG');
insert into table10 (code) values ('CDF');
select * from (
select 'ABC' as x
union all select 'BBB'
union all select 'TTT'
) t where t.x not in (select code from table10);
-- returns: BBB
-- TTT
See SQL Fiddle.
This can also be achieved using a stored procedure:
DELIMITER //
drop function if exists testcsv
//
create function testcsv(csv varchar(255)) returns varchar(255)
deterministic
begin
declare pos, found int default 0;
declare this, notin varchar(255);
declare continue handler for not found set found = 0;
set notin = '';
repeat
set pos = instr(csv, ',');
if (pos = 0) then
set this = trim('"' from csv);
set csv = '';
else
set this = trim('"' from trim(substring(csv, 1, pos-1)));
set csv = substring(csv, pos+1);
end if;
select 1 into found from table1 where code = this;
if (not found) then
if (notin = '') then
set notin = this;
else
set notin = concat(notin, ',', this);
end if;
end if;
until csv = ''
end repeat;
return (notin);
end
//
select testcsv('"ABC","BBB","TTT","DFG"')
Output:
BBB, TTT
I was wondering about the possibility to count the null columns of row in SQL, I have a table Customer that has nullable values, simply I want a query that return an int of the number of null columns for certain row(certain customer).
This method assigns a 1 or 0 for null columns, and adds them all together. Hopefully you don't have too many nullable columns to add up here...
SELECT
((CASE WHEN col1 IS NULL THEN 1 ELSE 0 END)
+ (CASE WHEN col2 IS NULL THEN 1 ELSE 0 END)
+ (CASE WHEN col3 IS NULL THEN 1 ELSE 0 END)
...
...
+ (CASE WHEN col10 IS NULL THEN 1 ELSE 0 END)) AS sum_of_nulls
FROM table
WHERE Customer=some_cust_id
Note, you can also do this perhaps a little more syntactically cleanly with IF() if your RDBMS supports it.
SELECT
(IF(col1 IS NULL, 1, 0)
+ IF(col2 IS NULL, 1, 0)
+ IF(col3 IS NULL, 1, 0)
...
...
+ IF(col10 IS NULL, 1, 0)) AS sum_of_nulls
FROM table
WHERE Customer=some_cust_id
I tested this pattern against a table and it appears to work properly.
My answer builds on Michael Berkowski's answer, but to avoid having to type out hundreds of column names, what I did was this:
Step 1: Get a list of all of the columns in your table
SELECT COLUMN_NAME FROM INFORMATION_SCHEMA.COLUMNS WHERE TABLE_NAME = 'myTable';
Step 2: Paste the list in Notepad++ (any editor that supports regular expression replacement will work). Then use this replacement pattern
Search:
^(.*)$
Replace:
\(CASE WHEN \1 IS NULL THEN 1 ELSE 0 END\) +
Step 3: Prepend SELECT identityColumnName, and change the very last + to AS NullCount FROM myTable and optionally add an ORDER BY...
SELECT
identityColumnName,
(CASE WHEN column001 IS NULL THEN 1 ELSE 0 END) +
-- ...
(CASE WHEN column200 IS NULL THEN 1 ELSE 0 END) AS NullCount
FROM
myTable
ORDER BY
NullCount DESC
For ORACLE-DBMS only.
You can use the NVL2 function:
NVL2( string1, value_if_not_null, value_if_null )
Here is a select with a similiar approach as Michael Berkowski suggested:
SELECT (NVL2(col1, 0, 1)
+ NVL2(col2, 0, 1)
+ NVL2(col3, 0, 1)
...
...
+ NVL2(col10, 0, 1)
) AS sum_of_nulls
FROM table
WHERE Customer=some_cust_id
A more generic approach would be to write a PL/SQL-block and use dynamic SQL. You have to build a SELECT string with the NVL2 method from above for every column in the all_tab_columns of a specific table.
Unfortunately, in a standard SQL statement you will have to enter each column you want to test, to test all programatically you could use T-SQL. A word of warning though, ensure you are working with genuine NULLS, you can have blank stored values that the database will not recognise as a true NULL (I know this sounds strange).
You can avoid this by capturing the blank values and the NULLS in a statement like this:
CASE WHEN col1 & '' = '' THEN 1 ELSE 0 END
Or in some databases such as Oracle (not sure if there are any others) you would use:
CASE WHEN col1 || '' = '' THEN 1 ELSE 0 END
You don't state RDBMS. For SQL Server 2008...
SELECT CustomerId,
(SELECT COUNT(*) - COUNT(C)
FROM (VALUES(CAST(Col1 AS SQL_VARIANT)),
(Col2),
/*....*/
(Col9),
(Col10)) T(C)) AS NumberOfNulls
FROM Customer
Depending on what you want to do, and if you ignore mavens, and if you use SQL Server 2012, you could to it another way. .
The total number of candidate columns ("slots") must be known.
1. Select all the known "slots" column by column (they're known).
2. Unpivot that result to get a
table with one row per original column. This works because the null columns don't
unpivot, and you know all the column names.
3. Count(*) the result to get the number of non-nulls;
subtract from that to get your answer.
Like this, for 4 "seats" in a car
select 'empty seats' = 4 - count(*)
from
(
select carId, seat1,seat2,seat3,seat4 from cars where carId = #carId
) carSpec
unpivot (FieldValue FOR seat in ([seat1],[seat2],[seat3],[seat4])) AS results
This is useful if you may need to do more later than just count the number of non-null columns, as it gives you a way to manipulate the columns as a set too.
This will give you the number of columns which are not null. you can apply this appropriately
SELECT ISNULL(COUNT(col1),'') + ISNULL(COUNT(col2),'') +ISNULL(COUNT(col3),'')
FROM TABLENAME
WHERE ID=1
The below script gives you the NULL value count within a row i.e. how many columns do not have values.
{SELECT
*,
(SELECT COUNT(*)
FROM (VALUES (Tab.Col1)
,(Tab.Col2)
,(Tab.Col3)
,(Tab.Col4)) InnerTab(Col)
WHERE Col IS NULL) NullColumnCount
FROM (VALUES(1,2,3,4)
,(NULL,2,NULL,4)
,(1,NULL,NULL,NULL)) Tab(Col1,Col2,Col3,Col4) }
Just to demonstrate I am using an inline table in my example.
Try to cast or convert all column values to a common type it will help you to compare the column of different type.
I haven't tested it yet, but I'd try to do it using a PL\SQL function
CREATE OR REPLACE TYPE ANYARRAY AS TABLE OF ANYDATA
;
CREATE OR REPLACE Function COUNT_NULL
( ARR IN ANYARRAY )
RETURN number
IS
cnumber number ;
BEGIN
for i in 1 .. ARR.count loop
if ARR(i).column_value is null then
cnumber := cnumber + 1;
end if;
end loop;
RETURN cnumber;
EXCEPTION
WHEN OTHERS THEN
raise_application_error
(-20001,'An error was encountered - '
||SQLCODE||' -ERROR- '||SQLERRM);
END
;
Then use it in a select query like this
CREATE TABLE TEST (A NUMBER, B NUMBER, C NUMBER);
INSERT INTO TEST (NULL,NULL,NULL);
INSERT INTO TEST (1 ,NULL,NULL);
INSERT INTO TEST (1 ,2 ,NULL);
INSERT INTO TEST (1 ,2 ,3 );
SELECT ROWNUM,COUNT_NULL(A,B,C) AS NULL_COUNT FROM TEST;
Expected output
ROWNUM | NULL_COUNT
-------+-----------
1 | 3
2 | 2
3 | 1
4 | 0
This is how i tried
CREATE TABLE #temptablelocal (id int NOT NULL, column1 varchar(10) NULL, column2 varchar(10) NULL, column3 varchar(10) NULL, column4 varchar(10) NULL, column5 varchar(10) NULL, column6 varchar(10) NULL);
INSERT INTO #temptablelocal
VALUES (1,
NULL,
'a',
NULL,
'b',
NULL,
'c')
SELECT *
FROM #temptablelocal
WHERE id =1
SELECT count(1) countnull
FROM
(SELECT a.ID,
b.column_title,
column_val = CASE b.column_title
WHEN 'column1' THEN a.column1
WHEN 'column2' THEN a.column2
WHEN 'column3' THEN a.column3
WHEN 'column4' THEN a.column4
WHEN 'column5' THEN a.column5
WHEN 'column6' THEN a.column6
END
FROM
( SELECT id,
column1,
column2,
column3,
column4,
column5,
column6
FROM #temptablelocal
WHERE id =1 ) a
CROSS JOIN
( SELECT 'column1'
UNION ALL SELECT 'column2'
UNION ALL SELECT 'column3'
UNION ALL SELECT 'column4'
UNION ALL SELECT 'column5'
UNION ALL SELECT 'column6' ) b (column_title) ) AS pop WHERE column_val IS NULL
DROP TABLE #temptablelocal
Similary, but dynamically:
drop table if exists myschema.table_with_nulls;
create table myschema.table_with_nulls as
select
n1::integer,
n2::integer,
n3::integer,
n4::integer,
c1::character varying,
c2::character varying,
c3::character varying,
c4::character varying
from
(
values
(1,2,3,4,'a','b','c','d'),
(1,2,3,null,'a','b','c',null),
(1,2,null,null,'a','b',null,null),
(1,null,null,null,'a',null,null,null)
) as test_records(n1, n2, n3, n4, c1, c2, c3, c4);
drop function if exists myschema.count_nulls(varchar,varchar);
create function myschema.count_nulls(schemaname varchar, tablename varchar) returns void as
$BODY$
declare
calc varchar;
sqlstring varchar;
begin
select
array_to_string(array_agg('(' || trim(column_name) || ' is null)::integer'),' + ')
into
calc
from
information_schema.columns
where
table_schema in ('myschema')
and table_name in ('table_with_nulls');
sqlstring = 'create temp view count_nulls as select *, ' || calc || '::integer as count_nulls from myschema.table_with_nulls';
execute sqlstring;
return;
end;
$BODY$ LANGUAGE plpgsql STRICT;
select * from myschema.count_nulls('myschema'::varchar,'table_with_nulls'::varchar);
select
*
from
count_nulls;
Though I see that I didn't finish parametising the function.
My answer builds on Drew Chapin's answer, but with changes to get the result using a single script:
use <add_database_here>;
Declare #val Varchar(MAX);
Select #val = COALESCE(#val + str, str) From
(SELECT
'(CASE WHEN '+COLUMN_NAME+' IS NULL THEN 1 ELSE 0 END) +' str
FROM INFORMATION_SCHEMA.COLUMNS WHERE TABLE_NAME = '<add table name here>'
) t1 -- getting column names and adding the case when to replace NULLs for zeros or ones
Select #val = SUBSTRING(#val,1,LEN(#val) - 1) -- removing trailling add sign
Select #val = 'SELECT <add_identity_column_here>, ' + #val + ' AS NullCount FROM <add table name here>' -- adding the 'select' for the column identity, the 'alias' for the null count column, and the 'from'
EXEC (#val) --executing the resulting sql
With ORACLE:
Number_of_columns - json_value( json_array( comma separated list of columns ), '$.size()' ) from your_table
json_array will build an array with only the non null columns and the json_query expression will give you the size of the array
There isn't a straightforward way of doing so like there would be with counting rows. Basically, you have to enumerate all the columns that might be null in one expression.
So for a table with possibly null columns a, b, c, you could do this:
SELECT key_column, COALESCE(a,0) + COALESCE(b,0) + COALESCE(c,0) null_col_count
FROM my_table
I am trying to create a delimitted string from the results of a query in DB2 on the iSeries (AS/400). I've done this in T-SQL, but can't find a way to do it here.
Here is my code in T-SQL. I'm looking for an equivelant in DB2.
DECLARE #a VARCHAR(1000)
SELECT #a = COALESCE(#a + ', ' + [Description], [Description])
FROM AP.Checkbooks
SELECT #a
If the descriptions in my table look like this:
Desc 1
Desc 2
Desc 3
Then it will return this:
Desc 1, Desc 2, Desc 3
Essentially you're looking for the equivalent of MySQL's GROUP_CONCAT aggregate function in DB2. According to one thread I found, you can mimic this behaviour by going through the XMLAGG function:
create table t1 (num int, color varchar(10));
insert into t1 values (1,'red'), (1,'black'), (2,'red'), (2,'yellow'), (2,'green');
select num,
substr( xmlserialize( xmlagg( xmltext( concat( ', ', color ) ) ) as varchar( 1024 ) ), 3 )
from t1
group by num;
This would return
1 red,black
2 red,yellow,green
(or should, if I'm reading things correctly)
You can do this using common table expressions (CTEs) and recursion.
with
cte1 as
(select description, row_number() over() as row_nbr from checkbooks),
cte2 (list, cnt, cnt_max) AS
(SELECT VARCHAR('', 32000), 0, count(description) FROM cte1
UNION ALL
SELECT
-- No comma before the first description
case when cte2.list = '' THEN RTRIM(CHAR(cte1.description))
else cte2.list || ', ' || RTRIM(CHAR(cte1.description)) end,
cte2.cnt + 1,
cte2.cnt_max
FROM cte1,cte2
WHERE cte1.row_nbr = cte2.cnt + 1 AND cte2.cnt < cte2.cnt_max ),
cte3 as
(select list from cte2
where cte2.cnt = cte2.cnt_max fetch first 1 row only)
select list from cte3;
I'm trying to do this in OLEDB and from what I understand you can't do this because you can't do anything fancy in SQL for OLEDB like declare variables or create a table. So I guess there is no way.
If you are running DB2 9.7 or higher, you can use LISTAGG function. Have a look here:
http://pic.dhe.ibm.com/infocenter/db2luw/v9r7/index.jsp?topic=%2Fcom.ibm.db2.luw.sql.ref.doc%2Fdoc%2Fr0058709.html