Unusual date conversion

[chandra.shekhar@tcs.com]

Hi,
My source is a File and contains a varchar field which actually stores date in the basic format YYYY-MM-DD.
In the transformer I didn't use any conversion methods for date conversion since the date is in default datastage format. Well this worked perfectly until I validated the data.
For spaces, some random date e.g 2009-07-28, 2010-05-04 etc is getting generated. What could be the reason for this ?

[ray.wurlod]

The only explanation is that the design specifies generation of a random date. Does the extended column metadata include a random Generator property?

[chandra.shekhar@tcs.com]

The extended column metadata doesn't have anything enabled..

[ray.wurlod]

Then you'd better show us exactly what you are doing.

[chandra.shekhar@tcs.com]

That's what Ray, I am doing nothing extra.
Just directly mapping a String Field(which contains date) to a Date Column.
Ideally I should use a StringToDate function but since my dates(from source) are in default YYYY-MM-DD format so I didn't use any conversion.

[prasannakumarkk]

Why dont you give trim(varahcar column).
is it a fixed lenght or delimited file?

[chulett]

I would advise to never let a Parallel job an implicit conversion like that, especially with a flat file source. Test for validity, explicitly convert the good dates and handle the bad or missing ones.

[chandra.shekhar@tcs.com]

I am now handling spaces in my job and explicitly converting into dates.
Datastage is really a mysterious tool

[bob7027]

You are right ... btw what version you are using?

[chandra.shekhar@tcs.com]

@Bob,
We have two clients 8.1 and 8.7, and I have checked it in both.