HASH.HELP stops

[Precious]

Hi All,

Happy 2007!

I have been running tests in regards to a big hashed file that we are working with. [45+ million rows, key=varchar(20), data=integer]
I have used various configurations with varying results. As part of the testing, I have created the file as a dynamic hashed file with default settings, dynamic hashed file with settings based on results from the ANALYZE.FILE command.

In this process, I also created this hashed file as a static hashed file.
However, I am unable to check the file as HASH.HELP stops. No errors are reported, just 'Program HASH.HELP stopped'.

Any reason that this is happening?

Thanks,

[ray.wurlod]

'Program HASH.HELP stopped'.

Any reason that this is happening?

Probably. Report it to your support provider.

I have certainly used HASH.HELP on hashed files with more rows than 45 million, but they were small rows. I suspect that some kind of arithmetic overflow has occurred within the HASH.HELP program.

Is your static hashed file anywhere near 2GB in size?

[Precious]

Is your static hashed file anywhere near 2GB in size?

Yes. I guess that's the problem then.

Just out of interest. Are static hashed files bigger than dynamic hashed files.
I was able to create and test the same file as a dynamic hashed file without any issues, or was it just pure luck?

Thanks,

[ray.wurlod]

Just luck.

Dynamic hashed files store only the primary group buffers in one file, using a separate file for secondary buffers. This can mean (though don't rely on it) that more than 2GB could be stored in a dynamic hashed file.

The total size on disk should be comparable in either case; the only difference that can't be attributed to distributional variation (that is, your choice of hashing algorithm for the static hashed file) would be the second header (2KB) in a dynamic hashed file.

[Precious]

Thank you very much.