Reading from a sequential file stage

devidotcom · Post by **devidotcom** » Sun Nov 04, 2007 11:50 pm

Hi,

I am trying to read data from a sequential file stage and get the following error... any idea why..

Error Calling subroutine DSD.Browse (Action=5 ); check DataStage is correctly set up in the project EDQS_PROTO

We have many server jobs created in this project already.
Saw a similar post on this error but had a oracle stage as the input from where the data was being read. The solution was to have a user-defined query.
But in this case where we ahve a sequential file as input stage what is the solution.

Thanks...

ArndW · Post by **ArndW** » Sun Nov 04, 2007 11:55 pm

If you simplify the job and read from the sequential stage and write to a sequential stage (use /dev/null as the output file) do you still get the same error?

Oh, wait "DSD.Browse" -> this might be a corrupt internal metadata hashed file. There have been a lot of threads recently regarding broken projects and hashed files - could your site have been among them?

devidotcom · Post by **devidotcom** » Mon Nov 05, 2007 2:45 am

thanks for the reply. That could be the reason. I tried reading the file in another project it worked fine.

I have another issue with the same file. The file has 1000 records which are being read with a server job with pipe delimited, quote char as double, default padding as # and missing column action as Pad with SQL Null.All the records gets read properly with the server job using a sequential file stage.

The same file with parallel job reads 21 records less the format table as delimiter as |, final delimiter as end and quote as double.

I put a reject link to the sequential file stage to see the 21 records rejected. One of the record is

""00999"|"001"|"93912"|"POBX.1"|"15555"|"000"|"00001"|"""THE CALIFORNIAN"""|"123 W ALISAL ST"|"PO BOX 81091"|"SALINAS"|"CA"|"N"|"939121091"|||"000000000"|"00000"|"06"|"053"|"0018"|"0018"|"00"|"2005-09-30"||"B900"|"91"|"5""

This record is being read in server job but not in parallel job. It can be seen that the last value 5 has two double quotes. How is it working in server job. Also I get a warning in the parallel job due to which it rejects that record
Field "DP_ADDR_CHECK_DGT" with 'delim=end' did not consume entire input, at offset: 212 [impexp/group_comp.C:6169]

devidotcom · Post by **devidotcom** » Mon Nov 05, 2007 9:19 pm

Hi All,

could not find a solution to this problem. Is it that the parallel jobs are more stringent about the data formats that are received when reading that data?

ArndW · Post by **ArndW** » Mon Nov 05, 2007 9:29 pm

devidotcom - yes, PX is quite picky about data types. But you haven't described your problems enough and therefore don't have responses.

Your example shows delimited data, but the error message looks like a fixed width issue. Until you fix your metadata you won't be able to do anything with this file.

ray.wurlod · Post by **ray.wurlod** » Mon Nov 05, 2007 10:29 pm

This error occurred during View Data. Does your job actually run?

I would suspect that the index on DS_METADATA has become out of synch. This table probably needs to be re-indexed. Schedule some time when no DataStage client is using the project and re-index the Repository.