Curious Date Behavior in Transformer

lshort · Post by **lshort** » Thu Jul 28, 2005 11:08 am

My job imports data from Seq file. Date format in file is "yyyymmdd". Datatype = Date. I changed the default date format for the job to 'yyyymmdd' as well. I also set the column Date format as 'yyyymmdd'. My target is a Dataset with Datatype = Date.

Despite all the above mentioned settings if passed straight through a Transformer the date comes out as 'yyyy-mm-dd'. If I remove the Transfomer stage the date is exported as 'yyyymmdd'.

Anyone know what's going on here?
Does a Transformer stage have some implicit transformation stuff?

ray.wurlod · Post by **ray.wurlod** » Thu Jul 28, 2005 9:16 pm

You may need to fiddle with your date picture. The default is %yyyy-%mm-%dd (with the delimiters). Try explicitly specifying %yyyy%mm%dd (without delimiters).

lshort · Post by **lshort** » Fri Jul 29, 2005 7:28 am

I have done just that. In each of the places mentioned. Still the data converts itself to %yyyy-%mm-%dd. Its bizare.

The reason I believe there is a problem is that any stage other than the transformer yields the correct result...and the transformer does not have a place to change the date format aside from a derivation expression.

ukyrvd · Post by **ukyrvd** » Fri Jul 29, 2005 9:01 am

I am just curious to know ..what is happening after transformer stage .. i.e, if you pass output of XFM (yyyy-mm-dd) to any other stage .. what you are seeing at the output end ?? is it automatically converting to yyyymmdd or yyyy-mm-dd?

thanks