Assuming these are coming from a text file, you need to create a schema.ini to force the JET engine (an MS product that Tableau uses to connect) to stop making assumptions.
Read up on this there: http://kb.tableausoftware.com/articles/knowledgebase/displaying-data-types-correctly
A schema.ini looks like this:
Col1="Labor Level A" Text
Col2="Labor Level B" Text
Col3="Labor Level C" Text
Col5="In Punch" DateTime
And so on. You will often find the schema.ini already exists in your working folder - you need to add the col descriptions.
You may also want to vote up my idea around schema.ini files - getting Tableau to do a better job of telling you there may be issues, and allowing you to automatically create one: http://community.tableau.com/ideas/1104
I am actually reading it from a excel file. All my data items are numeric, but tableau somehow read it as dimensions rather than measures. I fixed it though manually by converting them to measures. The problem I have is that first few values of several columns are null, and in those cases, tableau treats the entire column as null. Is this something that I could fix with schema.ini file ?
Format the entire column as text, then you can convert it to a number in Tableau.