Re: Reading schema information

On Mon, 25 May 2015 09:21:34 -0600, dmp <danap(at)ttc-cmc(dot)net> wrote:
> dmp wrote:
>> dmp wrote:
>>> Mansour Al Akeel wrote:
...
>>>> We have:
>>>>
>>>> IS_GENERATEDCOLUMN String => Indicates whether this is a generated
>>>> column
>>>>
>>>> YES --- if this a generated column
>>>> NO --- if this not a generated column
>>>> empty string --- if it cannot be determined whether this is a
>>>> generated column
>>>>
>>>>
>>>> My questions is, is this a bug ? if not, how can I obtain this
>>>> information about a column (if it's generated or not) ?
>>>>
>> Hello,
>>
>> I'm sorry, after testing, currently the given name IS_GENERATEDCOLUMN
is
>> given the error as you describe. The index of 23 can be used to obtain
>> the
>> results desired.
>>
>> Did you search the mailing list for a report or issue on this?
>
> Sorry for an earlier error.
>
> rs = dbMetaData.getColumns(tableMetaData.getCatalogName(1),
> tableMetaData.getSchemaName(1),
> tableMetaData.getTableName(1), "%");
>
> On further check,
> org.postgresql.Abstractjdbc2DatabaseMetaData.getColumns() does
> not implement the named Field IS_GENERATEDCOLUMN. As indicated using an
> index
> of 23 instead of the named Field should work.

Index 23 is IS_AUTOINCREMENT which does not have the same meaning as
IS_GENERATEDCOLUMN (index 24) which was added to the JDBC specification in
Java 7 / JDBC 4.1. The former is only for one specific column type: auto
increment (eg serial), while the latter is for all calculated or otherwise
generated columns.

Mark