Regards, Vik

You can do that in a number of ways. I see for example you can export the file to a SQL Server table and then use T-SQL to get the value using Execute SQL Task. Another option is to re-save it as Excel and then use Jet to query it (also a SQL way). If you want to sift through it you can read it row by row in a Script Task that encompasses File.IO routines. A Script Transformation can do that (and is similar to the Script Task). You need to look into specifics e.g. the file size, what you are looking for and what you are more proficient in.Arthur My Blog

Hello ArthurZ, Thanks for your reply, I do have 1.5 million reords and it has 1.2 GM file size, if i use sort tramsfomation it is killing my time , thank you, VikRegards, Vik

So are you getting a single distinct value or all the distinct values/records? I still recommend the SQL approach. You must be just fine using SELECT DISTINCT ...Arthur My Blog

Agree with AuthurZ - for reasonable specification db server 1.5 million rows are not a big data( BTW what is your row size) you can still can use T-SQL approach http://uk.linkedin.com/in/ramjaddu

Read the file in using a Flat File source, then use either an Aggregate, grouping by the columns you want to be distinct, or write your own asynchronous script to use a Hashtable to do the same, or use a 3rd party component. Talk to me now on