Hash File Seems Corrupted

[tcat2000]

I have a job that works correctly with a hash file. When I view data in the input tab... all data seems ok... all data seems ok in the output tab too.

I have another job that's not working with a hash file. All data seems ok in the input tab... but when I look into the output tab... column definitions are ok but I have garbage data... ie @KEY @ID @

Any suggestions?

Thanks

[ml]

1- check that you are using the same hash name.
2- check that you have the same metadata structure and the same column number.
3- first column should be a key.

If these dosen't help... delete the hash file and create another one.

good luck

[ray.wurlod]

@KEY, @ID and @ are entries from the file dictionary, they ought not to appear in the data portion of the hashed file. My guess is that someone has somehow copied the dictionary entries into the data part. These will have "D", "I", "PH" or "X" in field #1, so should be easily detected.

To see them

Code: Select all

SELECT F0, F1 FROM hashedfilename USING DICT VOC WHERE F1 IN ('D', 'I', 'PH', 'X');

To delete them, use the same FROM and WHERE clauses in a DELETE statement.

[Gokul]

Hi Ray,

The Query Given will only work for Account Hash file.
What if the Hash file is not an Account file,but is created on specific Path.

Gokul

[bryan]

hi gokul


please do a search for "Setfile"

If i remember it right, the syntax should be like
SETFILE hashfilepathwithname hashfilename OVERWRITING


this would put a pointer to ur hashed file in an outside directory in the VOC

[Sreenivasulu]

When the jobs were corrupted. We rebuilt the indexes then it worked fine.

Hope the same could work for hash file corruption