IDS Forum

Re: RE: error 201 when using functional index

Posted By: LYNKZ MIKE
Date: Friday, 27 November 2009, at 2:30 p.m.

In Response To: Re: RE: error 201 when using functional index (Jonathan Leffler)

Excellent Jonathan, you cleared my doubts. Inmediatly im going to open the PMR with this case.

Thank you very much.

> Hi Jonathan, thanks for SQLIDEBUG tip, look at what i found:
>

Well done - you're onto something - not sure what, but...

> The real erorr sis 206 and ISAM 111, what says the data server cant find
> the
> table, then shows error 201, then program finish with a dbclose
>

There's several lots of weird stuff happening - and I don't have a good
explanation for it.
However...

This is trace with functional indexes created. After trace without
> functional
> index.
>
> +14282
> +14283 C->S (102) Time: 2009-11-27 10:23:32.00000
> +14284 SQ_PREPARE
> +14285 # values: 3
> +14286 CMD.....: "select * from edc_det where cod_emp =? and cod_pto =? and
> num_edc =? order by cla_ent desc
> " [90]
> +14287 SQ_NDESCRIBE
> +14288 SQ_WANTDONE
> +14289 SQ_EOT
> +14290
> +14291 S->C (20) Time: 2009-11-27 10:23:32.00000
> +14292 SQ_ERR
> +14293 SQL error..........: -206
> +14294 ISAM/RSAM error....: -111
> +14295 Offset in statement: 23
> +14296 Error message......: "edc_det" [7]
> +14297 SQ_EOT
> +14298
>

This is the server saying it doesn't know what you mean by table edc_det.
Are you sure that it exists with that name? Is the table used earlier in
the program, by any chance? It is very odd.
Is there a chance that edc_det is a temp table that's been dropped?

> +14299 C->S (24) Time: 2009-11-27 10:23:32.00000
> +14300 SQ_PREPARE
> +14301 # values: 0
> +14302 CMD.....: "a_pobla_2001" [12]
> +14303 SQ_NDESCRIBE
> +14304 SQ_WANTDONE
> +14305 SQ_EOT
> +14306
> +14307 S->C (12) Time: 2009-11-27 10:23:32.00000
> +14308 SQ_ERR
> +14309 SQL error..........: -201
> +14310 ISAM/RSAM error....: 0
> +14311 Offset in statement: 1
> +14312 Error message......: "" [0]
> +14313 SQ_EOT
>

This code is sending the statement "a_pobla_2001" to the database server,
and not unreasonably it is saying "I don't understand you". There's a
problem in your I4GL not detecting the -206 error. As I mentioned once
before, have you got WHENEVER ERROR STOP in place, and if not, are you
religiously checking every SQL statement? And how did you create this bogus
statement?

> ---------------------------------------------------------
>
> TRACE WITHOUT FUNCTIONAL INDEX:
>
> +14282
> +14283 C->S (102) Time: 2009-11-27 10:40:02.00000
> +14284 SQ_PREPARE
> +14285 # values: 3
> +14286 CMD.....: "select * from edc_det where cod_emp =? and cod_pto =? and
> num_edc =? order by cla_ent desc
> " [90]
> +14287 SQ_NDESCRIBE
> +14288 SQ_WANTDONE
> +14289 SQ_EOT
> +14290
> +14291 S->C (284) Time: 2009-11-27 10:40:02.00000
> +14292 SQ_DESCRIBE
> +14293 Stmt Type...........: 2
> +14294 Server Stmt Id......: 49
> +14295 Estimated Cost......: 0
> +14296 Size of output tuple: 64
> +14297 # output fields.....: 12
> +14298 Size of string table: 96
> +14299 0) Field 'cod_emp'
> +14300 Index into string table: 0
> +14301 Starting offset in tuple: 0
> +14302 Type....................: DECIMAL; NOT NULLABLE
> +14303 Length : 2048 (0x800)
> +14304
> +14305 1) Field 'cod_pto'
> +14306 Index into string table: 8
> +14307 Starting offset in tuple: 5
> +14308 Type....................: DECIMAL; NOT NULLABLE
> +14309 Length : 1024 (0x400)
>
> What you think??
>

Since it appears that you send the same statement at the same point in the
execution sequence, and the server does recognize the table when there is no
functional index and fails to recognize it when there is a functional index,
it is possible to think that there's a bug in the server that you are
somehow tickling. I'm a bit cautious about claiming 'server bug', but I
think you have enough evidence to go to IBM Informix Tech Support and
request assistance.

>
> On Fri, Nov 27, 2009 at 06:51, LYNKZ MIKE <yellr@telecom.com.co> wrote:
> > Hi Jonathan, thanks a lot for response, i typo the table name
> > The query is :
> >
> > select * from edcve where cod1=48
> >
> > We dont build the functional index neither the function for it in 4gl
> > application, we built them by dbaccess.
> >
> > When we drop the functional index the 4GL application does not show the
> > error
> > 201, then the functional index exists, the application shows error 201.
> >
> > And even is more rare that application stops with error when opens this
> > cursor:
> >
> > declare cursor1 cursor for select * from edc_det
> > where cod_emp = v_emp
> > and cod_pto = v_cod_pto
> > and num_edc = v_num_edc
> > order by cla_ent desc
> > end function
> >
>
> Wha't the 'end function' doing?
>
> > What it could be?? Its very strange,
> >
> > How do i do to use SQLIDEBUG=2 ??
> >
>
> export SQLIDEBUG=2:something
>
> (the something is used as a prefix to the dump file names).
>
> Then run:
>
> $ dbaccess dbname script
>
> $ minimal-i4gl-program
>
> You should have two files, both starting with 'something'.
>
> sqliprint something_...name1...
>
> sqliprint something_...name2...
>
> Look to see if there's anything startlingly different between what works
> (DB-Access) and what doesn't (I4GL). It may be uninformative - but since
> -201 syntax error often means a broken SQL statement, you may be able to
> spot what is wrong with what the I4GL sent to IDS.
>
> > > Hi Art, thanks a lot for response, i dont understand well, The 4GL
> > > application cant run queries on table with functional indexes??
> >
> > It can run the queries; it just has problems building the functional
> > indexes.
> >
> > > we create the functional indexes using dbaccess just once.
> >
> > I thought we were looking at I4GL code trying to create the functional
> > index.
> > If not, then my previous answer was less relevant (even 'less than
> > relevant').
> >
> > > The 4GL compiler syntax checks all SQL and it's parser doesn't
> understand
> > > functional indexes. It only supports 7.3x supported syntax. You can
> > prepare
> > > and execute the create index as a string or include it in a BEGIN
> SQL...
> > > END SQL block to bypass the parser.
> > >
> > > Art
> > >
> > > Subject: error 201 when using functional index [18210]
> > >
> > > Good afternoon, we are having a rare problem when we are using a
> > functional
> > > index. There is a 4GL which gives error 201 when there is create a
> > > functional
> > > index.
> > >
> > > We are using IDS 10.FC9 on AIX 5.3 and 4GL 7.32
> > >
> > > The 4GL application get broken if this functional index is created:
> > >
> > > create table "informix".edcve
> > > (
> > > cod1 decimal(8,0) not null ,
> > > cod2 decimal(4,0) not null
> > > );
> > >
> > > CREATE FUNCTION f_test1( cod1 decimal(8,0) ) RETURNS decimal(8,0)
> > > WITH (NOT VARIANT);
> > > RETURN cod1 ;
> > > END FUNCTION;
> > >
> > > create index ix_edcve on edcve(f_test1(cod1));
> > >
> > > update statistics for table edcve;
> > >
> > > Debugging and using the onmode -I we have found the 4GL stops with
> error
> > > 201
> > > when this sql is executed:
> > >
> > > select * from edc_det where cod_emp =? and cod_pto =? and num_edc =?
> > > order by cla_ent desc
> > >
> >
> > This does not reference the table ecdve with the functional index AFAICS.
> >
> > > When this index is not created the 4GL does not get broken.
> > > Before this sql statement the 4GL querys the table in the following
> way:
> > > select * from edcdev where cod1=48
> > >
> >
> > Neither does this query reference the table edcde; it is referencing
> > another
> > table edcdev.
> >
> > Are you sure the functional index is the factor?
> >
> > Ultimately, both DB-Access and I4GL use versions of ESQL/C to access the
> > database.
> >
> > In general, that means they should get similar results. However,
> DB-Access
> > doesn't allow you to play with placeholders where I4GL does, so they are
> > not
> > completely equivalent.
> >
> > Sadly, DB-Access uses a different version of ESQL/C (built at the same
> time
> > as the server) from the separate CSDK product. That means there could be
> a
> > bug fix in the DB-Access version that isn't in the I4GL version.
> >
> > If the problem is consistent, then I would probably want to look at the
> > information flowing between the client (DB-Access or I4GL) and the server
> -
> > using SQLIDEBUG=2:something in the environment to capture the
> > communications, and then using sqliprint on the log files to compare the
> > information flows. You'd want to keep the program/script very small - the
> > bare minimum that reproduces the problem.
> >
> > Typically, this isolates the problem to 'client-side' or 'server-side'.
> >
> > > Both queries dont get broken if are executed in dbaccess.
> > > Anyone knows why is this happening??
> >
> > No real idea - it shouldn't be happening.
>

As I noted - sqliprint can often help isolate the problem to 'client-side'
or 'server-side'.
From what you've shown of the two scenarios, you are getting two divergent
reactions from the server to the same requests from the client depending on
the presence or absence of the functional index - that usually means that
the server has a problem rather than the client.

Nevertheless, the client code trace also suggests you are not handling all
errors correctly (it may be just this one place, it may be more general -
that I can't tell).

--
Jonathan Leffler #include <disclaimer.h>
Email: jleffler@earthlink.net, jleffler@us.ibm.com
Guardian of DBD::Informix v2008.0513 -- http://dbi.perl.org/
"Blessed are we who can laugh at ourselves, for we shall never cease to be
amused."
NB: Please do not use this email for correspondence.
I don't necessarily read it every week, even.
Jonathan Swift<http://www.brainyquote.com/quotes/authors/j/jonathan_swift.html>
- "May you live every day of your life."

Messages In This Thread

• error 201 when using functional index
  LYNKZ MIKE -- Thursday, 26 November 2009, at 3:40 p.m.
  • RE: error 201 when using functional index
    Art S. Kagel -- Thursday, 26 November 2009, at 6:57 p.m.
    • Re: RE: error 201 when using functional index
      LYNKZ MIKE -- Friday, 27 November 2009, at 8:52 a.m.
      • Re: RE: error 201 when using functional index
        Jonathan Leffler -- Friday, 27 November 2009, at 9:25 a.m.
        • Re: RE: error 201 when using functional index
          LYNKZ MIKE -- Friday, 27 November 2009, at 9:51 a.m.
          • Re: RE: error 201 when using functional index
            Jonathan Leffler -- Friday, 27 November 2009, at 10:02 a.m.
            • Re: RE: error 201 when using functional index
              LYNKZ MIKE -- Friday, 27 November 2009, at 10:49 a.m.
              • Re: RE: error 201 when using functional index
                Jonathan Leffler -- Friday, 27 November 2009, at 12:40 p.m.
                • Re: RE: error 201 when using functional index
                  LYNKZ MIKE -- Friday, 27 November 2009, at 2:30 p.m.
      • Re: RE: error 201 when using functional index
        Art Kagel -- Friday, 27 November 2009, at 11:18 a.m.
        • Re: RE: error 201 when using functional index
          LYNKZ MIKE -- Friday, 27 November 2009, at 2:23 p.m.
          • Re: RE: error 201 when using functional index
            Art Kagel -- Friday, 27 November 2009, at 2:46 p.m.
  • Re: error 201 when using functional index
    Jonathan Leffler -- Friday, 27 November 2009, at 9:09 a.m.

IDS Forum is maintained by Administrator with WebBBS 5.12.