Re: BUG #11528: Max Index Keys

The following bug has been logged on the website:

Bug reference: 11528
Logged by: Paul Dasari
Email address: pdasari(at)mdsol(dot)com
PostgreSQL version: 9.3.4
Operating system: Windows 2008 server
Description:

We are using Amazon Redshift postgreSQL. no matter what jar file we use it
gives the following error message.

Upon browsing, found that the source code is reading the max_index_keys
should be greater than 0. However the redshift database does not require
indexing. I really appreciate your assistance if there is any specific jar
to use to avoid using the max_index_keys function while executing the query

Below is the error we encountered:

Unable to determine a value for MaxIndexKeys due to missing system catalog
data

Thanks
Paul

On 9/30/2014 11:44 AM, pdasari(at)mdsol(dot)com wrote:
> We are using Amazon Redshift postgreSQL. no matter what jar file we use it
> gives the following error message.

jar file? what is this jar file? Do you mean JDBC ? The
PostgreSQL JDBC driver is its own project, has its own mailling lists.
The bug report form you used is for the core database server, which has
no Java components.

> Upon browsing, found that the source code is reading the max_index_keys
> should be greater than 0. However the redshift database does not require
> indexing. I really appreciate your assistance if there is any specific jar
> to use to avoid using the max_index_keys function while executing the query
>
> Below is the error we encountered:
>
> Unable to determine a value for MaxIndexKeys due to missing system catalog
> data

While PostgreSQL doesn't require indexes, its standard practice in
relational databases to always define a primary key for every table.
I'm not sure why JDBC would be enforcing this.

--
john r pierce 37N 122W
somewhere on the middle of the left coast

Hi John,

Thanks much for your response. Yes I am using the JDBC Jar file from
supplied from
http://jdbc.postgresql.org/download/postgresql-8.4-703.jdbc4.jar

Could you please help me provide a link so that i can post my question to
the required support group.

Thanks
Paul

On Tue, Sep 30, 2014 at 3:10 PM, John R Pierce <pierce(at)hogranch(dot)com> wrote:

> On 9/30/2014 11:44 AM, pdasari(at)mdsol(dot)com wrote:
>
>> We are using Amazon Redshift postgreSQL. no matter what jar file we use it
>> gives the following error message.
>>
>
> jar file? what is this jar file? Do you mean JDBC ? The
> PostgreSQL JDBC driver is its own project, has its own mailling lists.
> The bug report form you used is for the core database server, which has no
> Java components.
>
> Upon browsing, found that the source code is reading the max_index_keys
>> should be greater than 0. However the redshift database does not require
>> indexing. I really appreciate your assistance if there is any specific jar
>> to use to avoid using the max_index_keys function while executing the
>> query
>>
>> Below is the error we encountered:
>>
>> Unable to determine a value for MaxIndexKeys due to missing system catalog
>> data
>>
>
> While PostgreSQL doesn't require indexes, its standard practice in
> relational databases to always define a primary key for every table.
> I'm not sure why JDBC would be enforcing this.
>
>
>
> --
> john r pierce 37N 122W
> somewhere on the middle of the left coast
>
>

On 9/30/2014 12:16 PM, Paul Dasari wrote:
> Thanks much for your response. Yes I am using the JDBC Jar file from
> supplied from
> http://jdbc.postgresql.org/download/postgresql-8.4-703.jdbc4.jar
>

you actually should be using
http://jdbc.postgresql.org/download/postgresql-9.3-1102.jdbc4.jar which
as the home page says, supports all versions of postgres newer than 7.2,
and JDK 1.6

> Could you please help me provide a link so that i can post my question
> to the required support group.

http://jdbc.postgresql.org/community/mailinglist.html

--
john r pierce 37N 122W
somewhere on the middle of the left coast

Thanks for the inputs with the mailing list.
No matter which version i use the error exists.
I don't know, it seems like a bug!

Thanks
Paul

On Tue, Sep 30, 2014 at 3:20 PM, John R Pierce <pierce(at)hogranch(dot)com> wrote:

> On 9/30/2014 12:16 PM, Paul Dasari wrote:
>
> Thanks much for your response. Yes I am using the JDBC Jar file from
> supplied from
> http://jdbc.postgresql.org/download/postgresql-8.4-703.jdbc4.jar
>
>
> you actually should be using
> http://jdbc.postgresql.org/download/postgresql-9.3-1102.jdbc4.jar
> which as the home page says, supports all versions of postgres newer than
> 7.2, and JDK 1.6
>
> Could you please help me provide a link so that i can post my question
> to the required support group.
>
>
> http://jdbc.postgresql.org/community/mailinglist.html
>
>
>
> --
> john r pierce 37N 122W
> somewhere on the middle of the left coast
>
>

On 9/30/2014 12:26 PM, Paul Dasari wrote:
> Thanks for the inputs with the mailing list.
> No matter which version i use the error exists.
> I don't know, it seems like a bug!
>

afaik, JDBC itself doesn't do this test, so what software are you
running in Java thats using JDBC ? Many Database GUI packages insist
on primary keys, otherwise they can't reliably identify the row to
delete or update in a 'data grid' style edit.

--
john r pierce 37N 122W
somewhere on the middle of the left coast

Ok. I am using SAP Business Objects - Information Design Tool.

Not sure if you have already knew about this page, however after looking at
this page i feel like a bug, because when i use the same JAR that came out
of this it perfectly worked fine

http://www.garrettpatterson.com/2013/08/13/connect-pentahomondrian-schema-workbench-to-amazon-redshift/

But would be good if it is coming directly from postgresql.

Thanks
Paul

On Tue, Sep 30, 2014 at 3:34 PM, John R Pierce <pierce(at)hogranch(dot)com> wrote:

> On 9/30/2014 12:26 PM, Paul Dasari wrote:
>
> Thanks for the inputs with the mailing list.
> No matter which version i use the error exists.
> I don't know, it seems like a bug!
>
>
> afaik, JDBC itself doesn't do this test, so what software are you running
> in Java thats using JDBC ? Many Database GUI packages insist on primary
> keys, otherwise they can't reliably identify the row to delete or update in
> a 'data grid' style edit.
>
>
>
> --
> john r pierce 37N 122W
> somewhere on the middle of the left coast
>
>

On 9/30/2014 12:55 PM, Paul Dasari wrote:
> Ok. I am using SAP Business Objects - Information Design Tool.
>
> Not sure if you have already knew about this page, however after
> looking at this page i feel like a bug, because when i use the same
> JAR that came out of this it perfectly worked fine
>
> http://www.garrettpatterson.com/2013/08/13/connect-pentahomondrian-schema-workbench-to-amazon-redshift/
>
> But would be good if it is coming directly from postgresql.
>

if its a Redshift specific bug, which that seems to imply, then you'll
need to take that up with Amazon, as Redshift is a fork of a rather old
and no longer supported version of PostgreSQL

this discussion really does belong over on the JDBC list.

--
john r pierce 37N 122W
somewhere on the middle of the left coast

John R Pierce <pierce(at)hogranch(dot)com> writes:
> if its a Redshift specific bug, which that seems to imply, then you'll
> need to take that up with Amazon, as Redshift is a fork of a rather old
> and no longer supported version of PostgreSQL

max_index_keys is a GUC variable that reflects a server build
parameter, namely the maximum number of columns allowed in an index.
I am guessing that the JDBC driver is trying to read that variable
and it's not working because Redshift is descended from a PG version
that predates whenever we added that GUC. (Which was a long time
ago :-(.) I have no idea *why* the JDBC driver would need to know that.

You really need to ask about this on the pgsql-jdbc mailing list,
not here. There may not be a good solution other than using an ancient
JDBC driver with Redshift ... but the people who would know read that
list.

regards, tom lane

Sure. Thanks Tom and John!
On Sep 30, 2014 4:44 PM, "Tom Lane" <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> John R Pierce <pierce(at)hogranch(dot)com> writes:
> > if its a Redshift specific bug, which that seems to imply, then you'll
> > need to take that up with Amazon, as Redshift is a fork of a rather old
> > and no longer supported version of PostgreSQL
>
> max_index_keys is a GUC variable that reflects a server build
> parameter, namely the maximum number of columns allowed in an index.
> I am guessing that the JDBC driver is trying to read that variable
> and it's not working because Redshift is descended from a PG version
> that predates whenever we added that GUC. (Which was a long time
> ago :-(.) I have no idea *why* the JDBC driver would need to know that.
>
> You really need to ask about this on the pgsql-jdbc mailing list,
> not here. There may not be a good solution other than using an ancient
> JDBC driver with Redshift ... but the people who would know read that
> list.
>
> regards, tom lane
>