NotificationsYou must be signed in to change notification settings
Fork6
Star31

Commit52b6053

committed

Fix tsmatchsel() to account properly for null rows.

ts_typanalyze.c computes MCE statistics as fractions of the non-null rows,which seems fairly reasonable, and anyway changing it in released versionswouldn't be a good idea. But then ts_selfuncs.c has to account for that.Failure to do so results in overestimates in columns with a significantfraction of null documents. Back-patch to 8.4 where this stuff wasintroduced.Jesper Krogh

1 parentde623f3 commit52b6053Copy full SHA for 52b6053

File tree

2 files changed

-0

lines changed

src
- backend/tsearch
  - ts_selfuncs.c
- include/catalog
  - pg_statistic.h

2 files changed

-0

lines changed

`‎src/backend/tsearch/ts_selfuncs.c`

Lines changed: 6 additions & 0 deletions

Original file line number	Diff line number	Diff line change
`@@ -189,11 +189,17 @@ tsquerysel(VariableStatData *vardata, Datum constval)`
`189`	`189`	`/* No most-common-elements info, so do without */`
`190`	`190`	`selec=tsquery_opr_selec_no_stats(query);`
`191`	`191`	`}`
	`192`	`+`
	`193`	`+/*`
	`194`	`+ * MCE stats count only non-null rows, so adjust for null rows.`
	`195`	`+ */`
	`196`	`+selec *= (1.0-stats->stanullfrac);`
`192`	`197`	`}`
`193`	`198`	`else`
`194`	`199`	`{`
`195`	`200`	`/* No stats at all, so do without */`
`196`	`201`	`selec=tsquery_opr_selec_no_stats(query);`
	`202`	`+/* we assume no nulls here, so no stanullfrac correction */`
`197`	`203`	`}`
`198`	`204`
`199`	`205`	`returnselec;`

`‎src/include/catalog/pg_statistic.h`

Lines changed: 2 additions & 0 deletions

Original file line number	Diff line number	Diff line change
`@@ -246,6 +246,8 @@ typedef FormData_pg_statistic *Form_pg_statistic;`
`246`	`246`	`* type with identifiable elements (for instance, tsvector). staop contains`
`247`	`247`	`* the equality operator appropriate to the element type. stavalues contains`
`248`	`248`	`* the most common element values, and stanumbers their frequencies. Unlike`
	`249`	`+ * MCV slots, frequencies are measured as the fraction of non-null rows the`
	`250`	`+ * element value appears in, not the frequency of all rows. Also unlike`
`249`	`251`	`* MCV slots, the values are sorted into order (to support binary search`
`250`	`252`	`* for a particular value). Since this puts the minimum and maximum`
`251`	`253`	`* frequencies at unpredictable spots in stanumbers, there are two extra`

0 commit comments

Comments

(0)

Movatterモバイル変換

Navigation Menu

Search code, repositories, users, issues, pull requests...

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Commit52b6053

File tree

2 files changed

2 files changed

`‎src/backend/tsearch/ts_selfuncs.c`

`‎src/include/catalog/pg_statistic.h`

0 commit comments