Dynamic SQL to get rows from information_schema - sql

Let’s say I’m looking for a specific column in my database so I have something like this
SELECT COLUMN_NAME, TABLE_NAME FROM INFORMATION_SCHEMA.COLUMNS WHERE COLUMN_NAME like ‘%employeeid%’
But I also want to know how many rows each table has, I was told I can do this using Dynamic SQL so I have this now
DECLARE
#tableName NVARCHAR(MAX),
#sql NVARCHAR(MAX),
#colName NVARCHAR(MAX);
DECLARE CUR_TABLE CURSOR FOR
SELECT TABLE_NAME FROM INFORMATION_SCHEMA.COLUMNS
OPEN CUR_TABLE
FETCH NEXT FROM CUR_TABLE
INTO #tableName
WHILE ##FETCH_STATUS = 0
BEGIN
SET #colName = '%employeeid%'
SET #sql = 'SELECT COLUMN_NAME, TABLE_NAME, (SELECT COUNT(*) FROM ' + #tableName +') AS ROWS FROM INFORMATION_SCHEMA.COLUMNS where column_name like ' + ''' + #colName + ''';
FETCH NEXT FROM CUR_TABLE
INTO #tableName
END;
CLOSE CUR_TABLE
DEALLOCATE CUR_TABLE
EXEC sp_executesql #sql
But this doesn't work, What I'm trying to do is query a table with the column I am looking for, with the table name, and number of rows in the table.
How can I fix this?

You can make use of SQL Server's dynamic management views to quickly obtain the row counts*.
Find all tables with a column named 'MyColumn' and their current rows:
select Schema_Name(t.schema_id) schemaName, t.name TableName, s.row_count
from sys.columns c
join sys.tables t on t.object_id = c.object_id
join sys.dm_db_partition_stats s on s.object_id = c.object_id and s.index_id <= 1
where c.name='MyColumn';
* Accurate except for frequently updated tables where there could be some lag

The following uses INFORMATION_SCHEMA, dynamic SQL, and STRING_AGG() to build a query that will return a single result set.
DECLARE #ColumnName sysname = 'ProductID'
DECLARE #Newline VARCHAR(2) = CHAR(13) + CHAR(10)
DECLARE #SqlTemplate NVARCHAR(MAX) =
+ 'SELECT'
+ ' ColumnName = <ColumnNameString>,'
+ ' TableName = <TableSchemaAndNameString>,'
+ ' Rows = (SELECT COUNT(*) FROM <TableSchemaAndName>)'
+ #Newline
DECLARE #UnionSql NVARCHAR(100) = 'UNION ALL ' + #Newline
DECLARE #Sql NVARCHAR(MAX) = (
SELECT STRING_AGG(
REPLACE(REPLACE(REPLACE(
#SqlTemplate
, '<ColumnNameString>', QUOTENAME(C.COLUMN_NAME, ''''))
, '<TableSchemaAndNameString>', QUOTENAME(C.TABLE_SCHEMA + '.' + C.TABLE_NAME, ''''))
, '<TableSchemaAndName>', QUOTENAME(C.TABLE_SCHEMA) + '.' + QUOTENAME(C.TABLE_NAME))
, #UnionSql)
WITHIN GROUP(ORDER BY C.TABLE_SCHEMA, C.TABLE_NAME)
FROM INFORMATION_SCHEMA.TABLES T
JOIN INFORMATION_SCHEMA.COLUMNS C
ON C.TABLE_SCHEMA = T.TABLE_SCHEMA
AND C.TABLE_NAME = T.TABLE_NAME
WHERE T.TABLE_TYPE = 'BASE TABLE' -- Omit views
AND C.COLUMN_NAME = #ColumnName
)
SET #Sql = #Sql + 'ORDER BY Rows DESC, TableName' + #Newline
--PRINT #Sql
EXEC (#Sql)
I generalized it a bit by adding TABLE_SCHEMA so that it could be used with the AdventureWorks database. See this db<>fiddle for a working demo. Also included is equivalent logic that uses FOR XML instead of STRING_AGG for older SQL Server versions.

Assuming that you are using SQL Server, here is a shorthand way using sp_msforeachtable.
DECLARE #ColumnName NVARCHAR(200) = 'ContactID'
CREATE TABLE #T
(
ColumnName NVARCHAR(200),
TableName NVARCHAR(200),
RecordCount INT
)
INSERT INTO #T (ColumnName, TableName)
SELECT
ColumnName = C.COLUMN_NAME,
TableName = '['+C.TABLE_SCHEMA+'].['+C.TABLE_NAME+']'
FROM
INFORMATION_SCHEMA.COLUMNS C
WHERE
C.COLUMN_NAME LIKE '%' + #ColumnName + '%'
EXEC SP_MSFOREACHTABLE 'IF EXISTS(SELECT * FROM #T WHERE TableName = ''?'') UPDATE #T SET RecordCount = (SELECT COUNT(*) FROM ? ) WHERE TableName = ''?'''
SELECT
ColumnName,TableName,
TableType = CASE
WHEN RecordCount IS NULL
THEN 'View'
ELSE 'Table'
END,
RecordCount
FROM
#T
ORDER BY
CASE WHEN RecordCount IS NULL THEN 'View' ELSE 'Table' END
DROP TABLE #T

Related

How to use loop for like or ilike function?

I am currently asked to validate the result of a global search which to find all records containing some keywords cross whole database. To do so, I need to check all the rows in each table that have keywords. The result for this global search is ready and partially of the result looks like the one below:
table_name
column_name
keyword
cnt
Wf_Process
Name_EN
FEC
11
Wf_Process
FTABLENAME
GB
14
ICCClass
Name_EN
GB
4
What I am trying to do is using the 'like' operator to extract all the data in each table where columns containing keywords. That is, I will use query such as:
select distinct Name_EN, FTABLENAME from Wf_Process where Name_EN like '%FEC%' or FTABLENAME like '%GB%'
and
select distinct Name_EN from ICCClass where Name_EN like '%GB%'
to pull out all data I need in whole database. If I only have three records, then it is not a problem. However my result returned over thousands tables have more than 10K rows containing keywords in total. Therefore, I was trying to use a loop to do this mission but I failed.
My question therefore is, does anyone have any idea to write a loop to do the 'like' search for all the tables in one time? Or is there another way rather than loop can do this in SQL Server?
Thank you very much!
You need dynamic SQL for this. You can generate a big UNION ALL query of all the tables together.
DECLARE #sql nvarchar(max) = (
SELECT STRING_AGG(CAST('
SELECT
' + QUOTENAME(con.table_name, '''') + ' table_name,
' + QUOTENAME(con.column_name, '''') + ' column_name,
' + QUOTENAME(con.keyword, '''') + ' keyword,
COUNT(*) cnt
FROM ' + QUOTENAME(s.name) + '.' + QUOTENAME(t.name) + ' t
WHERE t.' + QUOTENAME(c.name) + ' LIKE ' + QUOTENAME('%' + con.keyword + '%', '''')
AS nvarchar(max)), '
UNION ALL')
FROM YourConditions con
JOIN sys.tables t ON t.table_name
JOIN sys.schemas s ON s.schema_id = t.schema_id
JOIN sys.columns c ON c.object_id = t.object_id
WHERE t.name = con.table_name
AND c.name = con.column_name
);
PRINT #sql; -- your friend
EXEC sp_executesql #sql;
There are more efficient ways to do this if there are many columns or keywords per table, but this should get you started.
The question looks odd, thou can be done via dynamic sql, but not as a regular solution more like a one off or POC/test.
use tempdb
GO
drop table if exists search_result
go
create table search_result (
table_name sysname
,column_name sysname
,keyword varchar(100))
go
insert into search_result values
('Wf_Process', 'Name_EN', 'FEC')
,('Wf_Process' , 'FTABLENAME', 'GB')
,('ICCClass', 'Name_EN', 'GB')
GO
drop table if exists result
create table result (val varchar(500))
go
declare #col sysname
declare #tab sysname
declare #kw varchar(100)
declare #sql varchar(1000)
while exists (select * from search_result)
begin
select top 1
#tab = table_name
, #col = column_name
, #kw = keyword
from search_result
set #sql = concat('insert into result select ', #col, ' from ', #tab, ' where ', #col, ' like ''%', #kw, '%''' )
print(#sql)
exec (#sql)
delete from search_result
where table_name = #tab and column_name = #col and keyword = #kw
end
GO
select * from result
--initially trying to get all the column name with its respective table name, finding out if there is that particular keyword in all table,
--if yes, insert into temp table, if there aint keyword, it will insert the record too but the count of record will be zero.
--trying to make it dynamic and running it inside the loop by assigning row number to each table with different column.
IF OBJECT_ID('tempdb..#TEMP') IS NOT NULL
DROP TABLE #TEMP
CREATE TABLE #TEMP (TABLE_NAME NVARCHAR(200), COLUMN_NAME NVARCHAR(200), KEYWORD NVARCHAR(200), CNT INT)
DECLARE #TABLE NVARCHAR(200)
DECLARE #TABLE_ID INT
SET #TABLE_ID = 1
DECLARE #TABLE_NAME NVARCHAR (200)
DECLARE #FOR_LOOP INT
DECLARE #COLUMN NVARCHAR(200)
DECLARE #COLUMN_NAME NVARCHAR(200)
--TOTAL NO OF RECORDS FOR LOOP
SET #FOR_LOOP = (SELECT COUNT(*)
FROM SYS.tables T
JOIN SYS.all_columns C ON T.object_id = C.object_id
JOIN INFORMATION_SCHEMA.TABLES S ON S.TABLE_NAME = T.name)
DECLARE #STRINGS NVARCHAR(200)
SET #STRINGS = '%FEC%' --------->ENTER YOUR KEYWORD HERE, TRY ONE AT A TIME
DECLARE #COUNT INT
WHILE #TABLE_ID <= #FOR_LOOP
BEGIN
SET #TABLE = (SELECT CAST(TABLE_NAME AS NVARCHAR(200))
FROM
(SELECT ROW_NUMBER()OVER(ORDER BY T.NAME) TABLE_ID
,S.TABLE_SCHEMA SCHEMA_NAME ,T.NAME TABLE_NAME, C.name COLUMN_NAME
FROM SYS.tables T
JOIN SYS.all_columns C ON T.object_id = C.object_id
JOIN INFORMATION_SCHEMA.TABLES S ON S.TABLE_NAME = T.name)A
WHERE TABLE_ID = #TABLE_ID)
SET #TABLE_NAME = '['+#TABLE+']'
SET #COLUMN = (SELECT CAST(COLUMN_NAME AS NVARCHAR(200))
FROM
(SELECT ROW_NUMBER()OVER(ORDER BY T.NAME) TABLE_ID
,S.TABLE_SCHEMA SCHEMA_NAME ,T.NAME TABLE_NAME, C.name COLUMN_NAME
FROM SYS.tables T
JOIN SYS.all_columns C ON T.object_id = C.object_id
JOIN INFORMATION_SCHEMA.TABLES S ON S.TABLE_NAME = T.name)A
WHERE TABLE_ID = #TABLE_ID)
SET #COLUMN_NAME = '['+#COLUMN+']'
DECLARE #EXEC NVARCHAR(200) = 'SELECT ' + #COLUMN_NAME +' FROM ' + #TABLE_NAME + ' WHERE ' + #COLUMN_NAME + ' LIKE ' + ''''+#STRINGS+''''
--THERE MUST BE ANOTHER WAY TO REPLACE SP_EXECUTESQL FOR BETTER PERFORMANCE, AS IT WILL BE EXECUTED AS MANY TIMES AS THERE ARE RECORDS IN #FOR_LOOP
EXEC SP_EXECUTESQL #EXEC
SET #COUNT = (SELECT ##ROWCOUNT)
IF (SELECT ##ROWCOUNT) >=1
BEGIN
INSERT INTO #TEMP VALUES
(#TABLE_NAME, #COLUMN_NAME, #STRINGS, #COUNT)
SET #TABLE_ID = #TABLE_ID + 1
END
ELSE
--DONT KNOW WHY THIS ELSE PART IS NOT TOUCHED, WHEREAS I THINK IT SHOULD HAVE
BEGIN
PRINT 'RECORD NOT FOUND'
END
END
GO
--AFTER THE ABOVE BLOCK IS EXECUTED THEN TRY RUNNING BELOW ONE
SELECT *FROM #TEMP WHERE CNT>0
--THERE MAY BE MULTIPLE ERROS, MUST BE EDITED AND CAN MAKE IT AS A PROCEDURE TOO.

How to get Max Date Value of Date column in Multiple tables

For example I have 2 tables in Database.
Ex :
Table T :
declare #t table (name varchar(20),DOB date)
Insert into #t (name,DOB) values ('Mohan','2001-07-19')
Insert into #t (name,DOB) values ('Minu','1998-06-19')
Table : TT
declare #tt table (name varchar(20),DOB date)
Insert into #tt (name,DOB) values ('Raju','2010-07-19')
Insert into #tt (name,DOB) values ('Rani','2001-06-19')
Now I have a Query to get Table name and column names of multiple tables basing on Date type filter .
SELECT C.TABLE_SCHEMA, c.TABLE_NAME,c.COLUMN_NAME
FROM INFORMATION_SCHEMA.COLUMNS c
JOIN sys.objects o ON o.name = c.TABLE_NAME
WHERE o.type = 'U' AND C.DATA_TYPE = 'Datetime'
Output :
T_Schema T_name T_column
dbo T DOB
dbo TT DOB
But how can I get max Date of both tables like below output :
T_Schema T_name T_column Max_dt
dbo T DOB 2001-07-19
dbo TT DOB 2010-07-19
Suggest me the best way to achieve this.
Here is another option that doesn't use a cursor. I doubt it will be any better from a performance perspective because you still need a subquery for every row. But I really hate cursors. I also used the system tables instead of the information schema views as those can sometimes be a bit odd. https://sqlblog.org/2011/11/03/the-case-against-information_schema-views
declare #SQL nvarchar(max) = N''
select #SQL = #SQL +
N'select SCHEMA_NAME = ''' + QUOTENAME(s.name) + ''', TABLE_NAME = '''
+ QUOTENAME(o.name) + ''', COLUMN_NAME = '''
+ QUOTENAME(c.name) + ''', MaxDate = '
+ '(select MAX(' + QUOTENAME(c.name) + ') from ' + QUOTENAME(s.name) + '.' + QUOTENAME(o.name) + ') UNION ALL '
from sys.columns c
join sys.systypes st on st.type = c.system_type_id
join sys.objects o on o.object_id = c.object_id and o.type = 'U'
join sys.schemas s on s.schema_id = o.schema_id
where st.name = 'datetime'
order by s.name
, o.name
, c.name
set #SQL = left(#SQL, len(#SQL) - 10) --removes final UNION ALL
select #SQL
--uncomment below when you are satisfied the dynamic sql is correct
--exec sp_executesql #SQL
Here's some dynamic SQL that should do what you want, via a cursor.
I'd caution using this if you have a lot of tables, or run this in test first. Cursors are not great performs generally speaking. You can run this against a system database, like master, which would have fewer values to see how it works.
create table #MaxDate (tname varchar(256), cname varchar(256), mdate datetime)
declare cur cursor local fast_forward
for
SELECT C.TABLE_SCHEMA, c.TABLE_NAME,c.COLUMN_NAME
FROM INFORMATION_SCHEMA.COLUMNS c
JOIN sys.objects o ON o.name = c.TABLE_NAME
WHERE o.type = 'U' AND C.DATA_TYPE = 'Datetime'
declare #schema varchar(64), #table varchar(256), #column varchar(256)
declare #sql varchar(max)
open cur
fetch next from cur into #schema, #table, #column
while ##FETCH_STATUS = 0
begin
set #sql = 'select ''' + #table + '''' + ',''' + '' + #column + '''' + ',' + 'max(' + #column + ') from ' + #schema + '.' + #table
print #sql
insert into #MaxDate
exec (#sql)
fetch next from cur into #schema, #table, #column
end
close cur
deallocate cur
select * from #MaxDate
drop table #MaxDate
Here's an answer using a cursor, dynamic SQL and a temporary table:
DECLARE table_cursor CURSOR LOCAL FAST_FORWARD FOR
SELECT
C.TABLE_SCHEMA,
c.TABLE_NAME,
c.COLUMN_NAME
FROM INFORMATION_SCHEMA.COLUMNS c
JOIN sys.objects o
ON o.name = c.TABLE_NAME
WHERE o.type = 'U'
AND C.DATA_TYPE = 'Datetime'
DECLARE #schema SYSNAME
DECLARE #table SYSNAME
DECLARE #column SYSNAME
DECLARE #sql NVARCHAR(1000)
CREATE TABLE #Data (SchemaName SYSNAME, TableName SYSNAME, ColumnName SYSNAME, MaxDate DATETIME)
OPEN table_cursor
FETCH NEXT FROM table_cursor INTO #schema, #table, #column
WHILE ##FETCH_STATUS = 0
BEGIN
SET #sql = 'INSERT INTO #Data (SchemaName, TableName, ColumnName, MaxDate) SELECT '''+#schema+''', '''+#table+''', '''+#column+''', MAX(['+#column+']) FROM ['+#schema+'].['+#table+']'
EXEC sp_executesql #sql
FETCH NEXT FROM table_cursor INTO #schema, #table, #column
END
CLOSE table_cursor
DEALLOCATE table_cursor
SELECT * FROM #Data
DROP TABLE #Data
SQL to BUILD SQL for you
DECLARE #SQL as nvarchar(max) ='';
select #SQL = #SQL + 'SELECT ' + Column_Name + ' adate, ''' + Column_Name + ''' colname, ''' + Table_name + ''' tabname FROM ' + Table_name + ' UNION ' FROM INFORMATION_SCHEMA.COLUMNS where data_type like '%date%'
select #SQL = 'SELECT TOP 100 * FROM (' + LEFT(#SQL, LEN(#SQL) -6) + ') IQ WHERE IQ.adate IS NOT null ORDER BY IQ.adate DESC';
--cut n paste the sql below, see what it does for you
select #SQL
mark II - executes sql for you, and sorts out names with spaces in them
DECLARE #SQL as nvarchar(max) ='';
select #SQL = #SQL + 'SELECT [' + Column_Name + '] adate, ''' + Column_Name + ''' colname, ''' + Table_name + ''' tabname FROM [' + Table_name + '] UNION ' FROM INFORMATION_SCHEMA.COLUMNS where data_type like '%date%'
select #SQL = 'SELECT TOP 100 * FROM (' + LEFT(#SQL, LEN(#SQL) -6) + ') IQ WHERE IQ.adate IS NOT null ORDER BY IQ.adate DESC';
select #SQL;
EXEC sp_executesql #sql;

Max(ID) of each table on a database

I am trying to get the maximum value MAX(ID) for each table I have which contains ID on my DB "Table_Example" and one schema_name in specific.
A single example:
SELECT MAX(ID) FROM Schema_name.Table_name1
This retrieve the maximum ID value that is located on Table_name1, but I have 84 tables. I would like to know the max of each table only in one column.
This is the code where I am working on currently:
I am using information_schema.columns to get the names of the tables automatic and the schema each table belongs to in order to get the whole DB IDs max(id) in one column.
USE TABLE_EXAMPLE
GO
DECLARE #ID NVARCHAR(MAX) --int
SET #ID = (SELECT DISTINCT COLUMN_NAME FROM INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_SCHEMA = 'SCHEMA_NAME' AND COLUMN_NAME IN ('ID') AND DATA_TYPE = 'INT')
SELECT #ID FROM (SELECT ('SCHEMA_NAME'+'.'+TABLE_NAME) AS TABLE_NAME FROM INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_SCHEMA = 'SCHEMA_NAME' AND COLUMN_NAME = 'ID' AND DATA_TYPE='INT') AS W
This Script retrieve wrong data but I think I am a bit closed to get the values, but I am not sure what I am doing wrong.
Could someone give me any good approach? Or any better option to get it done?
If you are wanting the max value in your identity columns, regardless of the names of those columns, then this is a very simple way of doing it. This will give you the Table Name, the name of the Identity Column, and the max value of that column:
SELECT sys.tables.name AS [Table Name],
sys.identity_columns.name AS [Column Name],
last_value AS [Last Value]
FROM sys.identity_columns
INNER JOIN sys.tables
ON sys.identity_columns.object_id = sys.tables.object_id
ORDER BY last_value DESC
This enumerate all tables with column Id and MAX value of this ID:
DECLARE #query nvarchar(MAX);
SELECT #query = COALESCE(#query + char(10)+'UNION ALL '+char(10)+'SELECT '''+QUOTENAME(s.name)+'.'+QUOTENAME(T.name)+''' [Table], MAX(Id) [Max] FROM '+QUOTENAME(s.name)+'.'+QUOTENAME(T.name),
'SELECT '''+QUOTENAME(s.name)+'.'+QUOTENAME(T.name)+''' [Table], MAX(Id) [Max] FROM '+QUOTENAME(s.name)+'.'+QUOTENAME(T.name))
FROM sys.schemas S
JOIN sys.tables T ON S.schema_id=T.schema_id
JOIN sys.columns C ON T.object_id=C.object_id
WHERE C.name='Id';
EXEC(#query);
Try like this,
This would give you the script.
SELECT DISTINCT 'SELECT MAX(' + + COLUMN_NAME + ') as ' + table_name + 'MaxId FROM ' + table_name
FROM INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_SCHEMA = 'dbo'
AND COLUMN_NAME IN ('ID')
CREATE TABLE #MaxValues (SchemaName SYSNAME , TableName SYSNAME , MaxID INT)
GO
Declare #SchemaName SYSNAME = 'dbo' --<-- Pass you schema name to this variable
,#ColumnName SYSNAME = 'ID' --<-- Column Name
,#DataType SYSNAME = 'INT' --<-- Data type
DECLARE #TableName SYSNAME , #SchmaName SYSNAME
, #Sql NVARCHAR(MAX) , #ColName SYSNAME;
Declare Cur CURSOR LOCAL FAST_FORWARD FOR
SELECT s.name , t.name , c.name
FROM sys.columns c
Inner join sys.tables t on c.object_id = t.object_id
Inner join sys.schemas s on s.schema_id = t.schema_id
Inner join sys.types tp on tp.user_type_id = c.user_type_id
WHERE s.name = #SchemaName
AND c.name = #ColumnName
AND tp.name = #DataType
OPEN Cur
FETCH NEXT FROM Cur INTO #SchmaName , #TableName , #ColName
WHILE (##FETCH_STATUS =0)
BEGIN
SET #Sql = N'INSERT INTO #MaxValues (SchemaName, TableName, MaxID )'
+ N' SELECT #SchmaName ,#TableName, MAX(' + QUOTENAME(#ColName) + N') '
+ N' FROM ' + QUOTENAME(#SchmaName) + '.' + QUOTENAME(#TableName)
Exec sp_executesql #Sql
,N'#SchmaName SYSNAME , #TableName SYSNAME'
,#SchmaName
,#TableName
FETCH NEXT FROM Cur INTO #SchmaName , #TableName , #ColName
END
CLOSE Cur
DEALLOCATE Cur
SELECT * FROM #MaxValues
Perhaps a little dynamic SQL
Edit This will return the Table Name(s) and Max ID in one dataset
Declare #SQL varchar(max) = '>>>'
Select #SQL = #SQL + SQL
From (
Select SQL='Union All Select TableName='''+concat('[',Table_Schema,'].[',Table_Name,']')+''',MaxID=max(ID) From '+concat('[',Table_Schema,'].[',Table_Name,'] ')
From INFORMATION_SCHEMA.COLUMNS
Where Column_Name = 'ID'
) A
Set #SQL=Replace(#SQL,'>>>Union All ','')
Exec(#SQL)
This script will list all the max ids. It is assuming your first column is the ID, regardless of its name.
DECLARE #Script AS VARCHAR(MAX) = ''
SELECT #Script = #Script + 'SELECT MAX(' + COLUMN_NAME + ') AS ID FROM ' + c.TABLE_NAME + ' UNION ALL ' + CHAR(13)+CHAR(10)
FROM INFORMATION_SCHEMA.COLUMNS c
INNER JOIN INFORMATION_SCHEMA.TABLES t ON c.TABLE_NAME = t.TABLE_NAME
WHERE c.ORDINAL_POSITION = 1 and t.TABLE_TYPE = 'BASE TABLE' and c.TABLE_SCHEMA = 'dbo' and c.DATA_TYPE = 'int'
SELECT #Script = LEFT(#Script, LEN(#Script) - 12)
EXEC (#Script)

SQL query to dynamically COUNT(FIELD) for all fields of table X

This should be such an easy thing, but it has me totally stumped.
You can easily return the count of each field of a table manually, with oneliners such as:
select count(FIELD1) from TABLE1 --42,706
select count(FIELD5) from TABLE1 --42,686
select count(FIELD9) from TABLE1 --2,918
This is slow and painful if you want to review several dozen tables the same way, and requires you to know the names of the fields in advance.
How handy would it be to have a script you can connect to any database, simply feed it a table name, and it will automatically return the counts for each field of that table?
Seems you can get half the work done with:
select COLUMN_NAME
from INFORMATION_SCHEMA.COLUMNS
where TABLE_NAME = 'TABLE1'
Something is flawed even with my barebones approach (explicitly hitting one field instead of them all):
declare #TABLENAME varchar(30), #FIELDNAME varchar(30)
set #TABLENAME = 'TABLE1'
set #FIELDNAME = (select top 1 COLUMN_NAME
from INFORMATION_SCHEMA.COLUMNS
where TABLE_NAME = #TABLENAME
and COLUMN_NAME = 'FIELD9')
select #FIELDNAME, count(#FIELDNAME) from TABLE1
The result is 42,706. Recall from my example above that FIELD9 only contains 2,918 values.
Even if that wasn't a problem, the more dynamic query would replace the last line with:
select #FIELDNAME, count(#FIELDNAME) from #TABLENAME
But SQL Server returns:
Must declare the table variable "#TABLENAME".
So I can avoid that by restructuring the query with a temp table:
declare #FIELDNAME varchar(30)
set #FIELDNAME = (select top 1 COLUMN_NAME
from INFORMATION_SCHEMA.COLUMNS
where TABLE_NAME = 'TABLE1'
and COLUMN_NAME = 'FIELD9')
if OBJECT_ID('TEMPDB..#TEMP1') is not null
drop table #TEMP1
select *
into #TEMP1
from TABLE1 --still not exactly dynamic!
select #FIELDNAME, count(#FIELDNAME) from #TEMP1
But that still brings us back to the original problem of returning 42,706 instead of 2,918.
I am running SQL Server 2008 R2, if it makes any difference.
Your query:
SELECT #FIELDNAME, COUNT(#FIELDNAME) FROM TABLE1
does not count FIELD9, #FIELDNAME is treated as a constant. It's like doing a COUNT(*).
You should use dynamic sql:
DECLARE #sql VARCHAR(MAX)
SET #sql = 'SELECT ''' + #fieldName + ''', COUNT([' + #fieldName + ']) FROM [' + #tableName + ']'
EXEC(#sql)
To get all columns and return it in a single result set without using a Temporary Table and CURSOR:
DECLARE #sql NVARCHAR(MAX) = ''
SELECT #sql = #sql +
'SELECT ''' + COLUMN_NAME + ''' AS ColName, COUNT([' + COLUMN_NAME + ']) FROM [' + #tableName + ']' + CHAR(10) +
'UNION ALL' + CHAR(10)
FROM INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_NAME = #tableName
SELECT #sql = LEFT(#sql, LEN(#sql) - 10)
EXEC(#sql)
Just set the #TargetTableName will do the job
DECLARE #TargetTableName sysname = '*'
SET NOCOUNT ON
DECLARE #TableName sysname, #ColumnName sysname, #Sql nvarchar(max)
DECLARE #TableAndColumn table
(
TableName sysname,
ColumnName sysname
)
DECLARE #Result table
(
TableName sysname,
ColumnName sysname,
NonNullRecords int
)
INSERT #TableAndColumn
SELECT o.name, c.name FROM sys.objects o INNER JOIN sys.columns c ON o.object_id = c.object_id
WHERE (o.name = #TargetTableName OR #TargetTableName = '*') AND o.type = 'U' AND c.system_type_id NOT IN (34, 35, 99) -- 34:image 35:text 99:ntext
ORDER BY c.column_id
DECLARE column_cursor CURSOR FOR SELECT TableName, ColumnName FROM #TableAndColumn
OPEN column_cursor
FETCH NEXT FROM column_cursor
INTO #TableName, #ColumnName
WHILE ##FETCH_STATUS = 0
BEGIN
SELECT #Sql = 'SELECT ''' + #TableName + ''' AS TableName, ''' + #ColumnName + ''' AS ColumnName, COUNT([' + #ColumnName + ']) AS NonNullRecords FROM [' + #TableName + ']'
print #Sql
INSERT #Result
EXEC (#Sql)
FETCH NEXT FROM column_cursor
INTO #TableName, #ColumnName
END
CLOSE column_cursor;
DEALLOCATE column_cursor;
SET NOCOUNT OFF
SELECT * FROM #Result

How to use EXEC or sp_executeSQL without looping in this case?

Environment: SQL Server 2005/2008,
pubs database
I have inserted into a table variable a set of data as shown below using information_schema tables.
Now I would like to update the flag column based on the result of executing the query in the column dSQL. I was able to update using loops/cursor and then used sp_executeSQL to
update the column and then update flag column later. But is there an alternate set-based way to do this without looping through all individual rows?
use pubs
go
declare #dsql Nvarchar(max)='', #tablename varchar(100), #colname varchar(100)
declare #t table (
TABLE_NAME varchar(100),
COLUMN_NAME varchar(100)
)
insert into #t
select distinct t.TABLE_NAME, c.COLUMN_NAME
from information_Schema.tables t
inner join
information_Schema.columns c
on t.TABLE_CATALOG = c.TABLE_CATALOG
where t.TABLE_SCHEMA = c.TABLE_SCHEMA
and t.TABLE_TYPE = 'BASE TABLE'
and c.DATA_TYPE = 'varchar'
select *, Dsql = 'select ' + COLUMN_NAME + ' from ' + TABLE_NAME + ' WHERE '
+ COLUMN_NAME + ' = ''Menlo Park''', '' as Flag
FROM #t
GO
I had an idea to create a function and call the function for each row to execute individual query statement but calling the function for each record might be a performance hit.
It's a loop or a function as you suggested (which is really a loop anyway).
Not possible, I made a script like it earlier.
declare #searchvalue varchar(100)
set nocount off
set #searchvalue = 'Hello world'
create table #tt (table_name varchar(64), column_name varchar(64), count int)
select * into #t from
(
select 'select ''' + a.table_name + ''' ''table_name'',''' + a.column_name + ''' ''column_name'', count(*) count from [' + a.table_name +'] where [' +a.column_name+']='''+#searchvalue +'''' + ' group by ['+ a.column_name+']' sqlstring
from INFORMATION_SCHEMA.COLUMNS a
join
INFORMATION_SCHEMA.TABLES b
on a.table_name = b.table_name
and b.table_type = 'base table'
where data_type = 'varchar'
) a
--loop cursor
Declare #sqlstring as nvarchar(500)
Declare SqlCursor CURSOR FAST_FORWARD FOR
SELECT sqlstring FROM #t
OPEN SqlCursor
FETCH NEXT FROM SqlCursor
INTO #sqlstring
WHILE ##FETCH_STATUS = 0
BEGIN
insert #tt
exec(#sqlstring)
FETCH NEXT FROM SqlCursor
INTO #sqlstring
END
CLOSE SqlCursor
DEALLOCATE SqlCursor
select * from #tt
drop table #tt
drop table #t
Use what you want
This is an old question, but I'd like to add a different answer all the same.
Try the following script (no cursor, no loop (according to execution plan)): (tested in MS SQL 2012)
-- Setting up test data/code
SELECT N'SELECT * FROM INFORMATION_SCHEMA.COLUMNS AS C' T
INTO #Code
UNION ALL
SELECT N'SELECT * FROM INFORMATION_SCHEMA.TABLES AS T'
-- Variable to hold the selected queries, seperated by CrLf. You can also add a "GO" or ";"
DECLARE #SQL NVARCHAR(MAX) = CHAR(13) + CHAR(10)
-- Concatenate the selected queries together into the variable
SELECT #SQL = #SQL + CHAR(13) + CHAR(10) + C.T
FROM #Code AS C
-- Execute
EXEC sys.sp_executesql #SQL
-- Clean up
DROP TABLE #Code