Gstat -table will only report one matching table

[mrotteveel]

When using gstat -table <name> ..., only the first matching table is reported, even if there are multiple matching tables in the database.

create database 'schemaexp.fdb' user sysdba;
create schema SCHEMA1;
create schema SCHEMA2;
create table SCHEMA1.TABLE1 (id integer);
create table SCHEMA1.TABLE2 (id integer);
create table SCHEMA2.TABLE2 (id integer);
exit;

Followed by

gstat -user sysdba -a -t TABLE2 schemaexp.fdb

Reports:

Gstat execution time Mon Oct 13 12:43:03 2025

Database header page information:
        Flags                   0
        Generation              13
        System Change Number    0
        Page size               8192
        ODS version             14.0
        Oldest transaction      9
        Oldest active           10
        Oldest snapshot         10
        Next transaction        10
        Next attachment ID      12
        Implementation          HW=AMD/Intel/x64 little-endian OS=Windows CC=MSVC
        Shadow count            0
        Page buffers            0
        Database dialect        3
        Database GUID:  {E0337DFD-A78A-4337-B39A-0973992CA582}
        Creation date           Oct 13, 2025 10:39:09
        Attributes              force write

    Variable header data:
        *END*

Analyzing database pages ...
TABLE2 (130)
    Primary pointer page: 289, Index root page: 290
    Pointer pages: 1, data page slots: 0
    Data pages: 0, average fill: 0%
    Primary pages: 0, secondary pages: 0, swept pages: 0
    Empty pages: 0, full pages: 0
    Fill distribution:
         0 - 19% = 0
        20 - 39% = 0
        40 - 59% = 0
        60 - 79% = 0
        80 - 99% = 0

Gstat completion time Mon Oct 13 12:43:04 2025

It should report both occurrences of TABLE2, as the -schema option was not specified.

Compare:

gstat -user sysdba -a -t TABLE2 -sch SCHEMA1 schemaexp.fdb

and

gstat -user sysdba -a -t TABLE2 -sch SCHEMA2 schemaexp.fdb

[mrotteveel]

Correction, judging by the table id, it seems to be the last matching table is reported

create database 'schemaexp.fdb' user sysdba;
create schema SCHEMA1;
create schema SCHEMA2;
create schema SCHEMA3;
create table SCHEMA1.TABLE1 (id integer);
create table SCHEMA1.TABLE2 (id integer);
create table SCHEMA2.TABLE2 (id integer);
create table SCHEMA3.TABLE2 (id integer);
exit;

Gstat execution time Mon Oct 13 12:58:29 2025

Database header page information:
        Flags                   0
        Generation              12
        System Change Number    0
        Page size               8192
        ODS version             14.0
        Oldest transaction      1
        Oldest active           2
        Oldest snapshot         2
        Next transaction        9
        Next attachment ID      3
        Implementation          HW=AMD/Intel/x64 little-endian OS=Windows CC=MSVC
        Shadow count            0
        Page buffers            0
        Database dialect        3
        Database GUID:  {B8FBDDF1-73B8-4387-80A4-6E7F34509E25}
        Creation date           Oct 13, 2025 10:58:18
        Attributes              force write

    Variable header data:
        *END*

Analyzing database pages ...
TABLE2 (131)
    Primary pointer page: 291, Index root page: 292
    Pointer pages: 1, data page slots: 0
    Data pages: 0, average fill: 0%
    Primary pages: 0, secondary pages: 0, swept pages: 0
    Empty pages: 0, full pages: 0
    Fill distribution:
         0 - 19% = 0
        20 - 39% = 0
        40 - 59% = 0
        60 - 79% = 0
        80 - 99% = 0

Gstat completion time Mon Oct 13 12:58:29 2025

[mrotteveel]

Similarly, I would expect

gstat -user sysdba -a -t TABLE1 -t TABLE2 -sch SCHEMA1 -sch SCHEMA2 schemaexp.fdb

to report the occurrence of TABLE2 in both schemas, but it only reports one (the one in SCHEMA2).

For comparison, run

gstat -user sysdba -a -t TABLE1 -t TABLE2 -sch SCHEMA1 schemaexp.fdb

[asfernandes]

These options will need to be changed, to -table schema.name and remove of -schema.
Parsing will be like ISQL -ch <charset>

[asfernandes]

These options will need to be changed, to -table schema.name and remove of -schema. Parsing will be like ISQL -ch

Forget this.

[pavel-zotov]

::: QA note :::
On Windows we can not use '-t ' to obtain statistics for non-ascii table names (neither as command line arg, nor as parameter to services API): exception 'table not found' raises in that case. No such problem on Linux.
Test currently uses ascii-only schema and table names.