gstat: Print detailed blob statistics and table size

[Noremos]

The new blob statistics looks as follow:

BLOBS (140)
    Primary pointer page: 753, Index root page: 754
    Total formats: 1, used formats: 1
    Average record length: 16.00, total records: 10
    Average version length: 0.00, total versions: 0, max versions: 0
    Average fragment length: 0.00, total fragments: 0, max fragments: 0
    Average unpacked length: 16.00, compression ratio: 1.00
    Pointer pages: 1, data page slots: 2
    Data pages: 2, average fill: 9%
    Primary pages: 1, secondary pages: 1, swept pages: 0
    Empty pages: 0, full pages: 0
    Blobs: 10, total length: 101451015, blob pages: 12439
        Level 0: 3, total length: 15, blob pages: 0
        Level 1: 6, total length: 1451000, blob pages: 182
        Level 2: 1, total length: 100000000, blob pages: 12257
    Table size: 101924864 bytes
    Fill distribution:
         0 - 19% = 2
        20 - 39% = 0
        40 - 59% = 0
        60 - 79% = 0
        80 - 99% = 0

The old one:

BLOBS (139)
    Primary pointer page: 319, Index root page: 320
    Total formats: 1, used formats: 1
    Average record length: 16.00, total records: 10
    Average version length: 0.00, total versions: 0, max versions: 0
    Average fragment length: 0.00, total fragments: 0, max fragments: 0
    Average unpacked length: 16.00, compression ratio: 1.00
    Pointer pages: 1, data page slots: 2
    Data pages: 2, average fill: 6%
    Primary pages: 1, secondary pages: 1, swept pages: 0
    Empty pages: 0, full pages: 0
    Blobs: 10, total length: 101451015, blob pages: 12257
        Level 0: 3, Level 1: 6, Level 2: 1
    Fill distribution:
         0 - 19% = 2
        20 - 39% = 0
        40 - 59% = 0
        60 - 79% = 0
        80 - 99% = 0

Gstat completion time Thu Jan 16 11:02:42 2025

[dyemanov]

Do we want it being backported into v5?

[hvlad]

Do we want it being backported into v5?

No objection

[sim1984]

You've messed something up with the formatting of the blob output by levels.

gstat -z
gstat version WI-V5.0.2.1597 Firebird 5.0 4a63159

gstat horses -u SYSDBA -p ***** -r -t HORSE

Analyzing database pages ...
HORSE (179)
    Primary pointer page: 709, Index root page: 710
    Total formats: 2, used formats: 1
    Average record length: 237.77, total records: 545113
    Average version length: 0.00, total versions: 0, max versions: 0
    Average fragment length: 0.00, total fragments: 0, max fragments: 0
    Average unpacked length: 2100.00, compression ratio: 8.83
    Pointer pages: 4, data page slots: 10176
    Data pages: 10176, average fill: 90%
    Primary pages: 9545, secondary pages: 631, swept pages: 0
    Empty pages: 7, full pages: 10165
    Blobs: 73761, total length: 7620692, blob pages: 0
        Level 0: 0, Level 1: 73761, Level 2: 7620692 <--- Wtf?
    Table size: 166789120 bytes
    Fill distribution:
         0 - 19% = 8
        20 - 39% = 0
        40 - 59% = 1
        60 - 79% = 2
        80 - 99% = 10165

[Noremos]

You've messed something up with the formatting of the blob output by levels.

gstat -z
gstat version WI-V5.0.2.1597 Firebird 5.0 4a63159

gstat horses -u SYSDBA -p ***** -r -t HORSE

Analyzing database pages ...
HORSE (179)
    Primary pointer page: 709, Index root page: 710
    Total formats: 2, used formats: 1
    Average record length: 237.77, total records: 545113
    Average version length: 0.00, total versions: 0, max versions: 0
    Average fragment length: 0.00, total fragments: 0, max fragments: 0
    Average unpacked length: 2100.00, compression ratio: 8.83
    Pointer pages: 4, data page slots: 10176
    Data pages: 10176, average fill: 90%
    Primary pages: 9545, secondary pages: 631, swept pages: 0
    Empty pages: 7, full pages: 10165
    Blobs: 73761, total length: 7620692, blob pages: 0
        Level 0: 0, Level 1: 73761, Level 2: 7620692 <--- Wtf?
    Table size: 166789120 bytes
    Fill distribution:
         0 - 19% = 8
        20 - 39% = 0
        40 - 59% = 1
        60 - 79% = 2
        80 - 99% = 10165

Sorry, forgot to change the message number

[Noremos]

There is one non-obvious behavior that i forgot to mention. Previously, blobs statistics were only displayed if they were present. I followed the same logic, so statistics for each level are displayed only if blobs of that level are present. Is this ok?

[sim1984]

There is one non-obvious behavior that i forgot to mention. Previously, blobs statistics were only displayed if they were present. I followed the same logic, so statistics for each level are displayed only if blobs of that level are present. Is this ok?

I think so. This may create a problem for tools interpreting statistics, but such problems always arise when changing the output format and they are quite easy to fix.

P.S. Maybe it is worth thinking about an alternative extensible format that does not break parsing when adding new data, for example JSON/XML. But this is for the future.