Hello,



Ok, I give up! Tried a lot of things in my code! But still get error on
index row size. So, I'm part of my code, if someone could help me! A valid
value for this type is: '(03-jan-2007 , 15-may-2010)'



typedef struct t_periodo

{

DateADT tvi;

DateADT tvf;

} Periodo;



Datum

periodo_in(PG_FUNCTION_ARGS)

{

char *str = PG_GETARG_CSTRING(0);

char tvi_char[MAXDATEFIELDS];

char tvf_char[MAXDATEFIELDS];



tvi_char = (char *) palloc(strlen(MAXDATEFIELDS));

tvf_char = (char *) palloc(strlen(MAXDATEFIELDS));



Periodo *result;



if (sscanf(str, " ( %s , %s )", tvi_char, tvf_char) != 2)

ereport(ERROR,

(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),

errmsg("invalid input syntax for periodo: \"%s\"", str)));



result->tvi = StringToDateADT(tvi_char);

result->tvi = StringToDateADT(tvf_char);



result = (Periodo *) palloc(sizeof(Periodo));



if (result->tvi > result->tvf)

ereport(ERROR,

(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),

errmsg("Initial date (TVi) must be smaller than final date
(TVf)")));



PG_RETURN_POINTER(result);

}





Please help me!

Thanks in advance!

Just glancing at this, a couple things stand out to me:

On Mon, 4 Jun 2007, Rodrigo Sakai wrote:

> Datum
> periodo_in(PG_FUNCTION_ARGS)
> {
> char *str = PG_GETARG_CSTRING(0);
> char tvi_char[MAXDATEFIELDS];
> char tvf_char[MAXDATEFIELDS];
>
> tvi_char = (char *) palloc(strlen(MAXDATEFIELDS));

What are you doing here? This is completely broken. I think you meant to
say:

char *tvi_char;

tvi_char = palloc(MAXDATEFIELDS);

Or:

char tvi_char[MAXDATEFIELDS];

and no palloc.

> tvf_char = (char *) palloc(strlen(MAXDATEFIELDS));

Same as above.

>
> Periodo *result;
>
> if (sscanf(str, " ( %s , %s )", tvi_char, tvf_char) != 2)

This is asking for trouble if arbitrary input can be fed to this.

> ereport(ERROR,
> (errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
> errmsg("invalid input syntax for periodo: \"%s\"", str)));
>
> result->tvi = StringToDateADT(tvi_char);
> result->tvi = StringToDateADT(tvf_char);
>
> result = (Periodo *) palloc(sizeof(Periodo));
>
> if (result->tvi > result->tvf)
> ereport(ERROR,
> (errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
> errmsg("Initial date (TVi) must be smaller than final date
> (TVf)")));
>
> PG_RETURN_POINTER(result);
> }
>
> Please help me!
>
> Thanks in advance!

Hope this helps.

--
My love, he's mad, and my love, he's fleet,
And a wild young wood-thing bore him!
The ways are fair to his roaming feet,
And the skies are sunlit for him.
As sharply sweet to my heart he seems
As the fragrance of acacia.
My own dear love, he is all my dreams --
And I wish he were in Asia.
-- Dorothy Parker

---------------------------(end of broadcast)---------------------------
TIP 6: explain analyze is your friend

Jeremy Drake <pgsql@jdrake.com> writes:
>> char tvi_char[MAXDATEFIELDS];
>>
>> tvi_char = (char *) palloc(strlen(MAXDATEFIELDS));

> What are you doing here? This is completely broken.

Indeed ... *please* tell us your compiler issued a warning about that.
If not an error --- none of the compilers I use will take it at all.
Maybe you need to get a real C compiler.

>> result->tvi = StringToDateADT(tvi_char);
>> result->tvi = StringToDateADT(tvf_char);
>>
>> result = (Periodo *) palloc(sizeof(Periodo));

Well, you oughta allocate result before you store into it, not after,
and I bet you meant to assign to the tvi and tvf fields, not tvi twice.

Bad as these mistakes are, they don't directly explain why a later index
entry creation would fail. What I'm betting is that your CREATE TYPE
command does not correctly describe the datatype size. Based on this
code it should be pass-by-reference, fixed-size-8-bytes, but I'll bet
a nickel your CREATE TYPE says something else --- probably varlena.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster

Hello Jeremy,

You are right about:
char tvi_char[MAXDATEFIELDS];

I have already tried this and didn't work too, it keeps giving the index
error!


about:
if (sscanf(str, " ( %s , %s )", tvi_char, tvf_char) != 2)

I test tvi_char and tvf_char with StringToDateADT(tvi_char). Just date
types are allowed!

Any other suggestions??
Thanks!

-----Original Message-----
From: Jeremy Drake [mailtogsql@jdrake.com]
Sent: segunda-feira, 4 de junho de 2007 00:23
To: Rodrigo Sakai
Cc: PostgreSQL Hackers
Subject: Re: [HACKERS] ERROR: index row size

Just glancing at this, a couple things stand out to me:

On Mon, 4 Jun 2007, Rodrigo Sakai wrote:

> Datum
> periodo_in(PG_FUNCTION_ARGS)
> {
> char *str = PG_GETARG_CSTRING(0);
> char tvi_char[MAXDATEFIELDS];
> char tvf_char[MAXDATEFIELDS];
>
> tvi_char = (char *) palloc(strlen(MAXDATEFIELDS));

What are you doing here? This is completely broken. I think you meant to
say:

char *tvi_char;

tvi_char = palloc(MAXDATEFIELDS);

Or:

char tvi_char[MAXDATEFIELDS];

and no palloc.

> tvf_char = (char *) palloc(strlen(MAXDATEFIELDS));

Same as above.

>
> Periodo *result;
>
> if (sscanf(str, " ( %s , %s )", tvi_char, tvf_char) != 2)

This is asking for trouble if arbitrary input can be fed to this.

> ereport(ERROR,
> (errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
> errmsg("invalid input syntax for periodo: \"%s\"", str)));
>
> result->tvi = StringToDateADT(tvi_char);
> result->tvi = StringToDateADT(tvf_char);
>
> result = (Periodo *) palloc(sizeof(Periodo));
>
> if (result->tvi > result->tvf)
> ereport(ERROR,
> (errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
> errmsg("Initial date (TVi) must be smaller than final date
> (TVf)")));
>
> PG_RETURN_POINTER(result);
> }
>
> Please help me!
>
> Thanks in advance!

Hope this helps.

--
My love, he's mad, and my love, he's fleet,
And a wild young wood-thing bore him!
The ways are fair to his roaming feet,
And the skies are sunlit for him.
As sharply sweet to my heart he seems
As the fragrance of acacia.
My own dear love, he is all my dreams --
And I wish he were in Asia.
-- Dorothy Parker


---------------------------(end of broadcast)---------------------------
TIP 7: You can help support the PostgreSQL project by donating at

http://www.postgresql.org/about/donate

>Indeed ... *please* tell us your compiler issued a warning about that.

I did shell scripts that suppress the warnings, so it was a big mistake, I
agree! But I think my problem is not here!

> Well, you oughta allocate result before you store into it, not after,
> and I bet you meant to assign to the tvi and tvf fields, not tvi twice.

You are right again. But this happened because I copy and paste the example
with a different order. But in my code all this is ok!

>Bad as these mistakes are, they don't directly explain why a later index
>entry creation would fail. What I'm betting is that your CREATE TYPE
>command does not correctly describe the datatype size. Based on this
>code it should be pass-by-reference, fixed-size-8-bytes, but I'll bet
>a nickel your CREATE TYPE says something else --- probably varlena.

Humm, this can help! I will investigate! Thanks a lot!


---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match

>but I'll bet
>a nickel your CREATE TYPE says something else --- probably varlena

You win, how can I pay this bet?
Very very thanks, I was looking for the error in the wrong place! It was
so simple and works!

But still have another problem, may be related with my output
functions!?!?
I still get an error while doing a select with order by clause! Any other
good bet?

The error is:

ERROR: cache lookup failed for operator......


Thanks in advance!


---------------------------(end of broadcast)---------------------------
TIP 4: Have you searched our list archives?

http://archives.postgresql.org