pbxt-discuss team mailing list archive
-
pbxt-discuss team
-
Mailing list archive
-
Message #00052
AVG_ROW_LENGTH computation
I'm trying to understand the computation listed on page 27 of the
downloaded slides from the MySQL Conference 2010.
Previously, it was recommended in a lp bug that "Average comp. rec. len"
should be used to save space and set as the AVG_ROW_LENGTH.
I get the part about "Byte size +1 of all columns".
How do you define the "Size of the Data log reference"?
Take this example; How can I obtain the optimal AVG_ROW_LENGTH to store
this data?
CREATE TABLE `audit` (
`ID` int(11) unsigned NOT NULL AUTO_INCREMENT,
`RefID` int(11) unsigned NOT NULL,
`AuditCatagory` varchar(64) NOT NULL,
`UserID` varchar(64) NOT NULL,
`Descript` text NOT NULL,
`EntryDate` date NOT NULL,
`EntryTime` time NOT NULL,
`Comments` text NOT NULL,
`Details` text NOT NULL,
PRIMARY KEY (`ID`),
KEY `db_index` (`RefID`) USING BTREE,
KEY `db_index_c` (`AuditCatagory`) USING BTREE
) ENGINE=PBXT AVG_ROW_LENGTH=124
CHECK TABLE: audit
Record buffer size = 176
Fixed length rec. len. = 174
Handle data record size = 138
Min/max header size = 14/26
Min/avg/max record size = 23/1407/196762
Avg row len set for tab = 124
Rows fixed length = NO
Maximum fixed size = 65536
Minimum variable size = 320
Minimum auto-increment = 2485698
Number of columns = 9
Number of fixed columns = 3
Columns req. for index = 3
Rec len req. for index = 4
Columns req. for blobs = 0
Number of blob columns = 0
Number of indices = 3
Extendend data length = 71393428
Minumum comp. rec. len. = 31
Average comp. rec. len. = 124
Maximum comp. rec. len. = 2316
Free record count = 2
Deleted record count = 0
Allocated record count = 2211916
Follow ups