Get number of rows for tables across all databases - sql

Question

Searched a bit and didn't see anything that really fit my needs (although I don't really understand SQL beyond the basic select statement so maybe I just missed it). Took a SQL class in school many moons ago.
I have:
Virtual Windows Server 2008 R2 running SQL 2008 R2. I connect via MS SQL Server Management Studio
90-100 databases
All DB's have the same table structure (each one is a different client)
I want to:
Search all databases and return a list of all databases which have a table (TableName) that is larger than say, 5000 records.
Some sort of script that I can schedule that will use that list and if it finds a DB that the TableName table is more than 5000 records, will delete anything older than x amount of days (say 30). Any kind of logging to know what happened over the last few days would be a bonus.
Any help would be appreciated. Thank you.
EDIT/UPDATE (2/24/15): Hiren Dhaduk provided a nice stored procedure that works. Thanks!

Answer 1

You can use following store procedure in any database. Then run this store procedure by passing table name and NoOfRows(Minimum number of rows for table , In your case it will be 5000):
CREATE PROCEDURE usp_FindLargeTables
#TableName VARCHAR(256) , #NoOfRows int
AS
BEGIN
DECLARE #DBName VARCHAR(256)
DECLARE #varSQL VARCHAR(512)
DECLARE #getDBName CURSOR
SET #getDBName = CURSOR FOR
SELECT name
FROM sys.databases
WHERE state != 6
CREATE TABLE #TmpTable (TableName VARCHAR(256),
SchemaName VARCHAR(256),
DBName VARCHAR(256))
OPEN #getDBName
FETCH NEXT
FROM #getDBName INTO #DBName
WHILE ##FETCH_STATUS = 0
BEGIN
SET #varSQL = 'USE ' + #DBName + ';
INSERT INTO #TmpTable
SELECT
sysobjects.Name as TableName
, sysindexes.Rows as NoOfRows , '''+ #DBName + ''' AS DBName
FROM
sysobjects
INNER JOIN sysindexes
ON sysobjects.id = sysindexes.id
WHERE
type = ''U''
AND sysindexes.IndId < 2
and sysobjects.Name = '''+ #TableName +''' and sysindexes.Rows > ' + CONVERT(VARCHAR, #NoOfRows) + ''
EXEC (#varSQL)
FETCH NEXT
FROM #getDBName INTO #DBName
END
CLOSE #getDBName
DEALLOCATE #getDBName
SELECT *
FROM #TmpTable
WHERE DBName != 'master'
-- STEP 2
DECLARE #DYNAMICQUERY VARCHAR(MAX)
SET #DYNAMICQUERY =
REPLACE((
SELECT 'DELETE FROM ['+ DBName +'].[dbo].['+ TableName +'] where Createdate < DATEADD(day, -30, GETDATE());'
FROM #TmpTable
WHERE DBName != 'master'
FOR XML PATH('')
), '<' , '<');
EXEC(#DYNAMICQUERY);
DROP TABLE #TmpTable
END
Example : usp_FindLargeTables 'DeltaStuds',5000
Clarification on your second point :
Unless you run a trace when the changes happen it is not possible.
So to do this i would suggest you to put one column called createdate in this table and then you will be able to delete record created before 30 days.

Answer 2

What you are asking is very custom to your setup...
You could use the following query to identify tables with row count greater than 5000
SELECT sc.name +'.'+ ta.name TableName
FROM sys.tables ta
INNER JOIN sys.partitions pa
ON pa.OBJECT_ID = ta.OBJECT_ID
INNER JOIN sys.schemas sc
ON ta.schema_id = sc.schema_id
WHERE ta.is_ms_shipped = 0 AND pa.index_id IN (1,0)
and ta.name = 'DeltaStuds'
GROUP BY sc.name,ta.name
having SUM(pa.rows)>5000
But you would need something like this link to run it for all DBs. How to find column names for all tables in all databases in SQL Server. I dont know of a easier way...
For the second part, you will have to setup a delete process that takes the resultset from the above query, creates a dynamic query and deletes rows based on your date field. In that process, you could inject an audit mechanism that logs that delete action, row count, datetime, etc...