27.06.2012 11:10, Thomas Steinmaurer wrote:
>
> gstat output has been changed/extended in Firebird 3. Can somebody
> comment on the new data returned by gstat? Or perhaps there is some
> description available in SVN.
Tables:
Primary pointer page: 179, Index root page: 180
Same as before.
Total formats: 1, used formats: 1
How many record formats are known totally (correspond to the number of
records in RDB$FORMATS for the given table) and how many of them are
actually used by existing record versions.
Average record length: 170.43, total records: 150000
Average version length: 0.00, total versions: 0, max versions: 0
Same as before.
Average fragment length: 0.00, total fragments: 0, max fragments: 0
Info about fragmented records (the ones that couldn't be put on the page
entirely and thus had to be splitted into multiple fragments).
Average unpacked length: 237.00, compression ratio: 1.39
Uncompressed (in memory) record length. Compression ratio = avg
uncompressed length / avg compressed length (reported above).
Pointer pages: 3, data page slots: 3915
How many pointer pages are allocated.
Data pages: 3915, average fill: 88%
Same as before.
Primary pages: 3915, full pages: 3914, swept pages: 0
3915 data pages have at least one primary (not back version or fragment)
record version. 3914 pages are marked as full (no more room). 0 pages
are known to have no garbage to collect.
Vlad will correct me if any mistakes were made in this chapter :-)
Indices:
Root page: 161017, depth: 2, leaf buckets: 107, nodes: 150000
Number of the index root page has been added.
Average node length: 5.72, total dup: 0, max dup: 0
Index node length has been added. It includes flag byte, prefix length,
suffix (aka data) length, suffix (aka data) itself, record number.
Average key length: 3.00, compression ratio: 1.28
How many bytes are occupied by keys only (node length - recno length).
How good is the compression (uncompressed length / compressed length).
Average prefix length: 2.83, average data length: 1.00
How long is the key prefix. Avg data length = same as before.
Clustering factor: 3915, ratio: 0.03
How many keys require a data page jump if accessing the record by its
recno, given that the keys are scanned sequentially. If it's close to
the data page count, then all records are physically stored in the index
order. If it's close to the record count, then records are randomly stored.
http://use-the-index-luke.com/glossary/index-clustering-factor
Dmitry
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
