From: Mark D. <mj...@ge...> - 2009-04-30 16:38:50
|
Mark Dominus wrote: > I will have it produce figures for the other tables. Quick summary: 27 GB. ============================================================================== Regenerating the statistics was taking too long, perhaps because of some sort of deadlock. But using the statistics from last week, here is a list of tables whose sizes exceed 1 MB: CHARSET_COLRANGE 216902 5639452 COLUMNRANGE 252922 9611036 DISCRETECHARSTATE 168094 10758016 MATRIXCOLUMN 5630858 310000000 MATRIXELEMENT 300000000 25800000000 MATRIXROW 182396 39032744 PHYLOCHAR 32405 1944300 PHYLOTREE 5603 1933035 PHYLOTREENODE 431761 53970124 SUB_TAXONLABEL 102749 2671474 TAXON 424785 33558016 TAXONLABEL 232262 16722864 TAXONLABELSET_TAXONLABEL 201327 6039810 TAXONVARIANT 548169 62491264 The second column is the number of records; the third is DB2's estimate of the total data size, excluding LOBs. To calculate this, it takes the average row size, adds ten bytes of overhead, and multiplies by the number of records. The row size includes the descriptor for any LOB fields but not the LOB data itself. To estimate LOB data, I got DB2 to give me a list of LOB fields; there were four. Then I did 'select sum(length(FIELD)) from TABLE'. The results are as follows: PHYLOTREE.NEWICKSTRING 6001170 STUDY_NEXUSFILE.NEXUS 333845436 HELP.HELPTEXT 24071 MATRIXROW.SYMBOLSTRING 307037386 Total size estimate exclusive of LOB data: 26 357.582 372 MB Total LOB estimate: 646.908 063 MB TOTAL ESTIMATED SIZE: 27 GB. |