Search for a string in all tables, rows and columns of a DB

Viewed 262870

I am lost in a big database and I am not able to find where the data I get comes from. I was wondering if it is possible with SQL Server 2005 to search for a string in all tables, rows and columns of a database?

Does anybody has an idea if it is possible and how?

15 Answers

This code should do it in SQL 2005, but a few caveats:

  1. It is RIDICULOUSLY slow. I tested it on a small database that I have with only a handful of tables and it took many minutes to complete. If your database is so big that you can't understand it then this will probably be unusable anyway.

  2. I wrote this off the cuff. I didn't put in any error handling and there might be some other sloppiness especially since I don't use cursors often. For example, I think there's a way to refresh the columns cursor instead of closing/deallocating/recreating it every time.

If you can't understand the database or don't know where stuff is coming from, then you should probably find someone who does. Even if you can find where the data is, it might be duplicated somewhere or there might be other aspects of the database that you don't understand. If no one in your company understands the database then you're in a pretty big mess.

DECLARE
    @search_string  VARCHAR(100),
    @table_name     SYSNAME,
    @table_schema   SYSNAME,
    @column_name    SYSNAME,
    @sql_string     VARCHAR(2000)

SET @search_string = 'Test'

DECLARE tables_cur CURSOR FOR SELECT TABLE_SCHEMA, TABLE_NAME FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_TYPE = 'BASE TABLE'

OPEN tables_cur

FETCH NEXT FROM tables_cur INTO @table_schema, @table_name

WHILE (@@FETCH_STATUS = 0)
BEGIN
    DECLARE columns_cur CURSOR FOR SELECT COLUMN_NAME FROM INFORMATION_SCHEMA.COLUMNS WHERE TABLE_SCHEMA = @table_schema AND TABLE_NAME = @table_name AND COLLATION_NAME IS NOT NULL  -- Only strings have this and they always have it

    OPEN columns_cur

    FETCH NEXT FROM columns_cur INTO @column_name
    WHILE (@@FETCH_STATUS = 0)
    BEGIN
        SET @sql_string = 'IF EXISTS (SELECT * FROM ' + QUOTENAME(@table_schema) + '.' + QUOTENAME(@table_name) + ' WHERE ' + QUOTENAME(@column_name) + ' LIKE ''%' + @search_string + '%'') PRINT ''' + QUOTENAME(@table_schema) + '.' + QUOTENAME(@table_name) + ', ' + QUOTENAME(@column_name) + ''''

        EXECUTE(@sql_string)

        FETCH NEXT FROM columns_cur INTO @column_name
    END

    CLOSE columns_cur

    DEALLOCATE columns_cur

    FETCH NEXT FROM tables_cur INTO @table_schema, @table_name
END

CLOSE tables_cur

DEALLOCATE tables_cur

If you are "getting data" from an application, the sensible thing would be to use the profiler and profile the database while running the application. Trace it, then search the results for that string.

The SSMS Tools PACK Add-In (Add-On) for Microsoft SQL Server Management Studio and Microsoft SQL Server Management Studio Express will do exactly what you need. On larger database it takes some time to search, but that is to be expected. It also includes a ton of cool features that should have be included with SQL Server Management Studio in the first place. Give it a try www.ssmstoolspack.com/

You do need to have SP2 for SQL Server Management Studio installed to run the tools.

I adapted a script originally written by Narayana Vyas Kondreddi in 2002. I changed the where clause to check text/ntext fields as well, by using patindex rather than like. I also changed the results table slightly. Unreasonably, I changed variable names, and aligned as I prefer (no disrespect to Mr. Kondretti). The user may want to change the data types searched. I used a global table to allow querying mid-processing, but a permanent table might be a smarter way to go.

/* original script by Narayana Vyas Kondreddi, 2002 */
/* adapted by Oliver Holloway, 2009 */

/* these lines can be replaced by use of input parameter for a proc */
declare @search_string varchar(1000);
set @search_string = 'what.you.are.searching.for';

/* create results table */
create table ##string_locations (
  table_name varchar(1000),
  field_name varchar(1000),
  field_value varchar(8000)
)
;
/* special settings */
set nocount on
;
/* declare variables */
declare
  @table_name varchar(1000),
  @field_name varchar(1000)
;
/* variable settings */
set @table_name = ''
;
set @search_string = QUOTENAME('%' + @search_string + '%','''')
;
/* for each table */
while @table_name is not null
begin

  set @field_name = ''
  set @table_name = (
    select MIN(QUOTENAME(table_schema) + '.' + QUOTENAME(table_name))
    from INFORMATION_SCHEMA.TABLES
    where 
      table_type = 'BASE TABLE' and
      QUOTENAME(table_schema) + '.' + QUOTENAME(table_name) > @table_name and
      OBJECTPROPERTY(OBJECT_ID(QUOTENAME(table_schema) + '.' + QUOTENAME(table_name)), 'IsMSShipped') = 0
  )

  /* for each string-ish field */
  while (@table_name is not null) and (@field_name is not null)
  begin
    set @field_name = (
      select MIN(QUOTENAME(column_name))
      from INFORMATION_SCHEMA.COLUMNS
      where 
        table_schema    = PARSENAME(@table_name, 2) and
        table_name  = PARSENAME(@table_name, 1) and
        data_type in ('char', 'varchar', 'nchar', 'nvarchar', 'text', 'ntext') and
        QUOTENAME(column_name) > @field_name
    )

    /* search that field for the string supplied */
    if @field_name is not null
    begin
      insert into ##string_locations
      exec(
        'select ''' + @table_name + ''',''' + @field_name + ''',' + @field_name + 
        'from ' + @table_name + ' (nolock) ' +
        'where patindex(' + @search_string + ',' + @field_name + ') > 0'  /* patindex works with char & text */
      )
    end
    ;
  end
  ;
end
;

/* return results */
select table_name, field_name, field_value from ##string_locations (nolock)
;
/* drop temp table */
--drop table ##string_locations
;

Actually Im agree with MikeW (+1) it's better to use profiler for this case.

Anyway, if you really need to grab all (n)varchar columns in db and make a search. See below. I suppose to use INFORMATION_SCHEMA.Tables + dynamic SQL. The plain search:

DECLARE @SearchText VARCHAR(100) 
SET @SearchText = '12'
DECLARE @Tables TABLE(N INT, TableName VARCHAR(100), ColumnNamesCSV VARCHAR(2000), SQL VARCHAR(4000))

INSERT INTO @Tables (TableName, ColumnNamesCSV)
SELECT  T.TABLE_NAME AS TableName, 
        ( SELECT C.Column_Name + ',' 
          FROM   INFORMATION_SCHEMA.Columns C 
          WHERE  T.TABLE_NAME = C.TABLE_NAME 
                 AND C.DATA_TYPE IN ('nvarchar','varchar') 
                 FOR XML PATH('')
        )
FROM    INFORMATION_SCHEMA.Tables T 

DELETE FROM @Tables WHERE ColumnNamesCSV IS NULL

INSERT INTO @Tables (N, TableName, ColumnNamesCSV)
SELECT ROW_NUMBER() OVER(ORDER BY TableName), TableName, ColumnNamesCSV  
FROM   @Tables

DELETE FROM @Tables WHERE N IS NULL

UPDATE @Tables 
SET ColumnNamesCSV = SUBSTRING(ColumnNamesCSV, 0, LEN(ColumnNamesCSV))

UPDATE @Tables 
SET SQL = 'SELECT * FROM ['+TableName+'] WHERE '''+@SearchText+''' IN ('+ColumnNamesCSV+')'

DECLARE @C INT, 
        @I INT, 
        @SQL VARCHAR(4000)

SELECT @I = 1, 
       @C = COUNT(1) 
FROM   @Tables

WHILE @I <= @C BEGIN
    SELECT @SQL = SQL FROM @Tables WHERE N = @I
    SET @I = @I+1
    EXEC(@SQL)
END

and one with LIKE clause:

DECLARE @SearchText VARCHAR(100) 
SET @SearchText = '12'

DECLARE @Tables TABLE(N INT, TableName VARCHAR(100), ColumnNamesCSVLike VARCHAR(2000), LIKESQL VARCHAR(4000))

INSERT INTO @Tables (TableName, ColumnNamesCSVLike)
SELECT   T.TABLE_NAME AS TableName, 
         (   SELECT  C.Column_Name + ' LIKE ''%'+@SearchText+'%'' OR ' 
             FROM    INFORMATION_SCHEMA.Columns C 
             WHERE   T.TABLE_NAME = C.TABLE_NAME 
                     AND C.DATA_TYPE IN ('nvarchar','varchar') 
          FOR XML PATH(''))
FROM     INFORMATION_SCHEMA.Tables T

DELETE FROM @Tables WHERE ColumnNamesCSVLike IS NULL

INSERT INTO @Tables (N, TableName, ColumnNamesCSVLike)
SELECT ROW_NUMBER() OVER(ORDER BY TableName), TableName, ColumnNamesCSVLike 
FROM @Tables

DELETE FROM @Tables WHERE N IS NULL

UPDATE @Tables 
SET  ColumnNamesCSVLike = SUBSTRING(ColumnNamesCSVLike, 0, LEN(ColumnNamesCSVLike)-2)

UPDATE @Tables SET LIKESQL = 'SELECT * FROM ['+TableName+'] WHERE '+ColumnNamesCSVLike

DECLARE @C INT, 
        @I INT, 
        @LIKESQL VARCHAR(4000)

SELECT @I = 1, 
       @C = COUNT(1) 
FROM @Tables

WHILE @I <= @C BEGIN
    SELECT @LIKESQL = LIKESQL FROM @Tables WHERE N = @I
    SET @I = @I +1
    EXEC(@LIKESQL)
END

To "find where the data I get comes from", you can start SQL Profiler, start your report or application, and you will see all the queries issued against your database.

This query can do the thing for you.

DECLARE
@search_string  VARCHAR(100),
@table_name     SYSNAME,
@table_id       INT,
@column_name    SYSNAME,
@sql_string     VARCHAR(2000)

SET @search_string = 'StringtoSearch'

DECLARE tables_cur CURSOR FOR SELECT ss.name +'.'+ so.name [name], object_id FROM sys.objects so INNER JOIN sys.schemas ss ON so.schema_id = ss.schema_id WHERE  type = 'U'

OPEN tables_cur

FETCH NEXT FROM tables_cur INTO @table_name, @table_id

WHILE (@@FETCH_STATUS = 0)
BEGIN
    DECLARE columns_cur CURSOR FOR SELECT name FROM sys.columns WHERE object_id = @table_id 
        AND system_type_id IN (167, 175, 231, 239, 99)

    OPEN columns_cur

    FETCH NEXT FROM columns_cur INTO @column_name
        WHILE (@@FETCH_STATUS = 0)
        BEGIN
            SET @sql_string = 'IF EXISTS (SELECT * FROM ' + @table_name + ' WHERE [' + @column_name + '] 
            LIKE ''%' + @search_string + '%'') PRINT ''' + @table_name + ', ' + @column_name + ''''

            EXECUTE(@sql_string)

        FETCH NEXT FROM columns_cur INTO @column_name
        END

    CLOSE columns_cur

DEALLOCATE columns_cur

FETCH NEXT FROM tables_cur INTO @table_name, @table_id
END

CLOSE tables_cur
DEALLOCATE tables_cur
Related