>>>>> "Tom" == Tom Lane <tgl@sss.pgh.pa.us> writes:
Tom> But we need to be clear in the documentation about what thisTom> property actually means. My objection to having
itanswer at theTom> index or column level is basically that that encourages confusionTom> as to what it means.
OK. Here's new output with the scope changes, and some of the names
changed in an attempt to make them clearer:
cap | AM | Index | Column
--------------------+----+-------+--------asc | | | tdesc | | |
fnulls_first | | | fnulls_last | | | torderable | | |
tdistance_orderable| | | freturnable | | | tsearch_array | | | tsearch_nulls
| | | tclusterable | | t | backward_scan | | t | index_scan | | t
|bitmap_scan | | t | can_order | t | | can_unique | t | | can_multi_col
| t | | can_exclude | t | |
(17 rows)
(The can_* names are reserved for the AM level where we're asking about
the abstract capabilities of the AM.)
Better? Worse? Any more improvements to the names?
--
Andrew (irc:RhodiumToad)