Hi, Valeria,

Thanks for noticing this change.  I should have made this part of the 
documentation and am forwarding this to the mailing list so that 
everyone is made aware of this.

To make the return from a select function suitable for select 
operation again, we need to name the columns in a way that avoids 
special characters.  Normally, the users who intend to do this should 
give aliases to the functions, e.g.,

select sum(in_bytes) as total_in, avg(out_bytes) as avg_out ...

However, if there is no alias provided, the internal generated names 
take on the form of function name followed by an integer, such as sum1 
and avg2.  The automatically inserted "count(*)" will appear as "count0".

Please consider using aliases if you expect the output to remain 
consistent.  If you have a suggestion on how the columns should be 
named, please let us know.

Thanks.

John


On 4/27/2010 1:56 AM, Valeria Lorenzetti wrote:
> Hello John,
> I was doing some tests and I noticed the change in column names when
> using aggregate functions like SUM, AVG, etc.
>
> In my program I used dumpNames() to print column names.
>
> Example of SELECT statment: "SUM(IN_BYTES),AVG(OUT_BYTES)"
>
> Column names printed by dumpNames():
> *sum1, avg2, count0*
> 0,247.5,2
>
> Expected output:
> *SUM(IN_BYTES), AVG(OUT_BYTES), count(*)*
> 0,247.5,2
>
> I'm pretty sure it was in such way in previous versions of the library.
> Not a big problem, but I wanted to tell you about this.
>
> Thanks,
> Valeria Lorenzetti
>
_______________________________________________
FastBit-users mailing list
[email protected]
https://hpcrdm.lbl.gov/cgi-bin/mailman/listinfo/fastbit-users

Reply via email to