hashed file stage

sravan · Post by **sravan** » Wed Oct 05, 2005 10:32 pm

Hi All,

Plz Clear me some doubt on Hashed file stage.

Can we use hashed file as source and Hashed file as a target?

thanks in advance

Sravan

tcj · Post by **tcj** » Wed Oct 05, 2005 10:48 pm

Hi Sravan,

Yes you can use hash files as a source as well as hash files as a target.

Or did you want to know if you could use them at the same time?

Tim

sravan · Post by **sravan** » Wed Oct 05, 2005 11:04 pm

Thanks tcj,

As well as clear me one more doubt. Is there any limit to use the particular no. of hashed files in the single job.

have a nice time

sravan

ray.wurlod · Post by **ray.wurlod** » Thu Oct 06, 2005 12:39 am

There's no theoretical limit.

However, your job might rapidly become unwieldy and difficult to maintain if there are too many stages - irrespective of their type.

What do you consider to be "many"? Why do you think you need that many? Is there a smarter solution?

tcj · Post by **tcj** » Thu Oct 06, 2005 12:41 am

Sravan,

I don't think there is a set limit to the number of hashed files that you can have in a single job.

The amount of hash files that you can have in your job will be limited by other factors. Such as the amount of RAM on the server. The design of your jobs. (I would be looking at you job design if you had over 10 hash files)

Cheers

Tim

kduke · Post by **kduke** » Thu Oct 06, 2005 5:47 am

You do not want a single hash file to be both a source and a target because of the way it processes hash files. You can do this in a job but not read transform and write to the same hash file. This is different than write to a hash file and then start reading from it.

DeepakCorning · Post by **DeepakCorning** » Thu Oct 06, 2005 2:05 pm

Hey want to add up a question along with this topic.
Is there any limit on length of data (or say volume of data ) which can stored in the hash file?
If i want to use a FILE as source and target then wt is better -- to use a seq file or hash file??
If it cannt handle any volume of data then wht is the max? And if go out of the limit wht error it wil give like unable to write to hash file the record ID etc etc.

kcbland · Post by **kcbland** » Thu Oct 06, 2005 2:12 pm

Upper limits for 64bits is high terabytes. Limit for 32 bit hash file is 2.2 gigabytes. DS only creates 32bit hash files thru the Designer, you have to use command line create statement to get a 64bit file.

LogicDude · Post by **LogicDude** » Thu Oct 06, 2005 3:46 pm

kcbland wrote:Upper limits for 64bits is high terabytes. Limit for 32 bit hash file is 2.2 gigabytes. DS only creates 32bit hash files thru the Designer, you have to use command line create statement to get a 64bit file.

Hi Ken,

I had no idea we can do this. I thought because Datastage is a 32 bit environment, the hashed files are always 32bit. Can you let us know more about the command line create statement to get a 64bit file? And the limitations.

Thanks a bunch !

ray.wurlod · Post by **ray.wurlod** » Thu Oct 06, 2005 5:13 pm

Fire up the Hashed File Calculator on your DataStage CD. It's in the Utilities\Unsupported folder. Play and learn!

LogicDude · Post by **LogicDude** » Thu Oct 06, 2005 6:15 pm

ray.wurlod wrote:Fire up the Hashed File Calculator on your DataStage CD. It's in the Utilities\Unsupported folder. Play and learn!

Thanks Ray. I will do that. And may be again get back to you if I have any concern.

I need to ask the client side people for the Datastage CD.