Re: Encoded Vector Index error

When you created the EVI how many unique values did you predict on the
CREATE EVI statement ?

It builds a (I believe) 2byte or 4byte binary field to contain the number
of unique values. Say it is built with a 2byte field,  if it grows to
contain more distinct values and the number is greater than the 2byte field
can contain it has to rebuild the entire thing using a new 4 byte field.
I had thought that this happened automatically though.

However, your error message sounds something like this situation.

I wonder why you built it on a file with such few number of records (8000+)
and not a file with a million or so where you would get the benefit of an
EVI.

Secondly,  I don't think a 50 byte field is a good candidate for an EVI.  I
would think a plain old index would serve you better.

EVI's I would think would better suited for States, Color codes, department
numbers, or other "Code" type fields with low cardinality(er,  number of
distinct values).

Don't tell me that out of 8000 records that you had 7500+  distinct values
now Peter?

Respectfully
John Carr


>
> Hi People,
> I have a FILEA with 8,178 records, which I attempted to copy into FILEB,
a
> brand new empty file which happened to have an EVI created over it, on a
> 50  character field containing last names. I got the following error:
>  Failure for device or member FILEB file FILEB in library LIBX.
> Checking the job log, I found these messages:
>  The maximum number of unique encoded vector index values has been
>    exceeded for data space index FILEBEVI FILEBEVI.

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---