Commit Graph

20051 Commits

Author SHA1 Message Date
Bruce Momjian
3b0a5e50d7 Update VACUUM VERBOSE FSM message, per Tom. 2005-04-24 03:51:49 +00:00
Tom Lane
35f9b461f1 Repair two TIME WITH TIME ZONE bugs found by Dennis Vshivkov. Comparison
of timetz values misbehaved in --enable-integer-datetime cases, and
EXTRACT(EPOCH) subtracted the zone instead of adding it in all cases.
Backpatch to all supported releases (except --enable-integer-datetime code
does not exist in 7.2).
2005-04-23 22:53:05 +00:00
Tom Lane
0e99be1c25 Remove useless argtype_inherit() code, and make consequent simplifications.
As I pointed out a few days ago, this code has failed to do anything useful
for some time ... and if we did want to revive the capability to select
functions by nearness of inheritance ancestry, this is the wrong place
and way to do it anyway.  The knowledge would need to go into
func_select_candidate() instead.  Perhaps someday someone will be motivated
to do that, but I am not today.
2005-04-23 22:09:58 +00:00
Bruce Momjian
ac8998f2e6 Turns out our existing page size is already optimal in most cases:
< * Research the use of larger page sizes
2005-04-23 21:45:28 +00:00
Bruce Momjian
f4dcb52efe Item already added to existing 'thread' item:
< * Consider parallel processing a single query
<
<   This would involve using multiple threads or processes to do optimization,
<   sorting, or execution of single query.  The major advantage of such a
<   feature would be to allow multiple CPUs to work together to process a
<   single query.
<
2005-04-23 21:44:52 +00:00
Bruce Momjian
6869683147 Remove item, not sure what it refers to:
< * Allow ORDER BY ... LIMIT 1 to select high/low value without sort or
<   index using a sequential scan for highest/lowest values
<
<   If only one value is needed, there is no need to sort the entire
<   table. Instead a sequential scan could get the matching value.
<
2005-04-23 21:43:24 +00:00
Bruce Momjian
1207d5b64e New item:
> * Change WAL to use 32-bit CRC, for performance reasons
2005-04-23 21:41:01 +00:00
Bruce Momjian
e847558031 Update threading item:
<   Solaris) might benefit from threading.
>   Solaris) might benefit from threading.  Also explore the idea of
>   a single session using multiple threads to execute a query faster.
2005-04-23 21:39:27 +00:00
Bruce Momjian
bb62899db5 Done:
< * Improve SMP performance on i386 machines
> * -Improve SMP performance on i386 machines
2005-04-23 21:38:42 +00:00
Tom Lane
9b5b9616f4 Remove explicit FreeExprContext calls during plan node shutdown. The
ExprContexts will be freed anyway when FreeExecutorState() is reached,
and letting that routine do the work is more efficient because it will
automatically free the ExprContexts in reverse creation order.  The
existing coding was effectively freeing them in exactly the worst
possible order, resulting in O(N^2) behavior inside list_delete_ptr,
which becomes highly visible in cases with a few thousand plan nodes.

ExecFreeExprContext is now effectively a no-op and could be removed,
but I left it in place in case we ever want to put it back to use.
2005-04-23 21:32:34 +00:00
Bruce Momjian
95c7bff47c Done:
< * Optimize locale to have minimal performance impact when not used
2005-04-23 21:25:49 +00:00
Bruce Momjian
9ff49665ce Fix typo:
< * Add ISo INTERVAL handling
> * Add ISO INTERVAL handling
2005-04-23 21:22:55 +00:00
Bruce Momjian
714d5a4c37 Update VACUUM VERBOSE update, per Alvaro. 2005-04-23 21:16:34 +00:00
Bruce Momjian
9ba6587f8b Update working of VACUUM VERBOSE. 2005-04-23 21:10:20 +00:00
Bruce Momjian
52e08c35f7 Make VACUUM VERBOSE FSM output all output in a single INFO output
statement.
2005-04-23 20:56:01 +00:00
Bruce Momjian
e8ad6d9666 Move info about lack of depencency checking in Makefiles to developer's faq.q 2005-04-23 20:52:32 +00:00
Bruce Momjian
f7e514d010 Update FAQ by eliminating non-frequent items like large objects and
extending questions.  Update wording of various entries.
2005-04-23 20:51:44 +00:00
Bruce Momjian
48e7a196ea Done:
> * -Allow non-bitmap indexes to be combined by creating bitmaps in memory
2005-04-23 19:13:40 +00:00
Bruce Momjian
97a4dad35f Update H4 tag to H3 to be consistent with heading levels. 2005-04-23 18:59:17 +00:00
Bruce Momjian
9f1e864d02 Add item about server-side debugging. 2005-04-23 18:57:46 +00:00
Bruce Momjian
c34ea747a6 Update FAQ items to point to existing web pages rather than duplication
such information.  Remove MySQL mention.  Move server-side debug item to
developer's FAQ.  Update URLs.
2005-04-23 18:57:25 +00:00
Tom Lane
19d127548c Add comment about checkpoint panic behavior during shutdown, per
suggestion from Qingqing Zhou.
2005-04-23 18:49:54 +00:00
Tom Lane
b1faf3624b Allow -2147483648 to be treated as an INT4 rather than INT8 constant.
Per discussion with Paul Edwards.
2005-04-23 18:35:12 +00:00
Tom Lane
d79eeef38b Treat a zero-D array as empty in int_enum(), per Andrew@supernews. 2005-04-23 17:55:50 +00:00
Tom Lane
3842892492 Recent changes got the sense of the notnull bit backwards in the 2.0
protocol output routines.  Mea culpa :-(.  Per report from Kris Jurka.
2005-04-23 17:45:35 +00:00
Tom Lane
c114e166e5 Define the right-hand input of AT TIME ZONE as a full a_expr instead of
c_expr.  Perhaps the restriction was once needed to avoid bison errors,
but it seems to work just fine now --- and even generates a slightly
smaller state machine.  This change allows examples like
SELECT '13:45'::timetz AT TIME ZONE '-07:00'::interval;
to work without parentheses around the right-hand input.
2005-04-23 17:22:16 +00:00
Bruce Momjian
0975b95bc3 Update FAQ release number to 8.0.2. 2005-04-23 15:22:56 +00:00
Bruce Momjian
e947e1153a Modify output of VACUUM VERBOSE to be clearer. 2005-04-23 15:20:39 +00:00
Tom Lane
0548497cc5 int_aggregate's int_enum() doesn't work correctly with arrays that
aren't 1-D, so give an error message instead of failing.  Per report
from Ron Mayer.
2005-04-23 05:38:22 +00:00
Tom Lane
56c8877291 Turns out that my recent elimination of the 'redundant' flatten_andors()
code in prepqual.c had a small drawback: the flatten_andors code was
able to cope with deeply nested AND/OR structures (like 10000 ORs in
a row), whereas eval_const_expressions tends to recurse until it
overruns the stack.  Revise eval_const_expressions so that it doesn't
choke on deeply nested ANDs or ORs.
2005-04-23 04:42:53 +00:00
Bruce Momjian
25434e32cb Add documentation stating how to determine if FSM settings are too low
--- look at a database-wide VACUUM VERBOSE.
2005-04-23 03:27:40 +00:00
Tom Lane
e092828241 Teach choose_bitmap_and() to actually be choosy --- that is, try to
make some estimate of which available indexes to AND together, rather
than blindly taking 'em all.  This could probably stand further
improvement, but it seems to do OK in simple tests.
2005-04-23 01:57:34 +00:00
Tom Lane
4b89126ccc Fix bogus EXPLAIN display of rowcount estimates for BitmapAnd and
BitmapOr nodes.
2005-04-23 01:29:15 +00:00
Tom Lane
bc843d3960 First cut at planner support for bitmap index scans. Lots to do yet,
but the code is basically working.  Along the way, rewrite the entire
approach to processing OR index conditions, and make it work in join
cases for the first time ever.  orindxpath.c is now basically obsolete,
but I left it in for the time being to allow easy comparison testing
against the old implementation.
2005-04-22 21:58:32 +00:00
Bruce Momjian
ccbb07d922 Fix typo:
<   Currently indexes do not have enough tuple tuple visibility
<   information to allow data to be pulled from the index without
<   also accessing the heap.  One way to allow this is to set a bit
<   to index tuples to indicate if a tuple is currently visible to
<   all transactions when the first valid heap lookup happens.  This
<   bit would have to be cleared when a heap tuple is expired.
>   Currently indexes do not have enough tuple visibility information
>   to allow data to be pulled from the index without also accessing
>   the heap.  One way to allow this is to set a bit to index tuples
>   to indicate if a tuple is currently visible to all transactions
>   when the first valid heap lookup happens.  This bit would have to
>   be cleared when a heap tuple is expired.
2005-04-22 15:40:16 +00:00
Bruce Momjian
6f61ddd40d Typo fix. Alvaro. 2005-04-22 15:32:58 +00:00
Bruce Momjian
8f4a1b3e84 Update URL for TODO list. 2005-04-22 13:38:19 +00:00
Bruce Momjian
d76f279a55 Remove pre-7.3 mention that FOR UPDATE can be before LIMIT.
Document that FOR UPDATE and LIMIT together can return fewer rows that
LIMIT specifies, and why.
2005-04-22 04:20:44 +00:00
Bruce Momjian
26bb65df1e Clarify that only crypt can't use md5 pg_shadow passwords. 2005-04-22 04:18:58 +00:00
Bruce Momjian
c82b895284 Clarify use of MD5 authentication and pg_shadow encryption. 2005-04-21 22:19:19 +00:00
Tom Lane
14c7fba3f7 Rethink original decision to use AND/OR Expr nodes to represent bitmap
logic operations during planning.  Seems cleaner to create two new Path
node types, instead --- this avoids duplication of cost-estimation code.
Also, create an enable_bitmapscan GUC parameter to control use of bitmap
plans.
2005-04-21 19:18:13 +00:00
Bruce Momjian
c6221db3c0 Updated text for bitmaps:
<   Bitmap indexes index single columns that can be combined with other bitmap
<   indexes to dynamically create a composite index to match a specific query.
<   Each index is a bitmap, and the bitmaps are bitwise AND'ed or OR'ed to be
<   combined.  They can index by tid or can be lossy requiring a scan of the
<   heap page to find matching rows, or perhaps use a mixed solution where
<   tids are recorded for pages with only a few matches and per-page bitmaps
<   are used for more dense pages.  Another idea is to use a 32-bit bitmap
<   for every page and set a bit based on the item number mod(32).

>   This feature allows separate indexes to be ANDed or ORed together.  This
>   is particularly useful for data warehousing applications that need to
>   query the database in an many permutations.  This feature scans an index
>   and creates an in-memory bitmap, and allows that bitmap to be combined
>   with other bitmap created in a similar way.  The bitmap can either index
>   all TIDs, or be lossy, meaning it records just page numbers and each
>   page tuple has to be checked for validity in a separate pass.
2005-04-21 15:20:39 +00:00
Bruce Momjian
631e03145f Done:
< * Add tool to query pg_stat_* tables and report indexes that aren't needed
<   or tables that might need indexes
2005-04-21 04:09:34 +00:00
Tom Lane
e6f7edb9d5 Install some slightly realistic cost estimation for bitmap index scans. 2005-04-21 02:28:02 +00:00
Tom Lane
2f8c7c866c Make pg_ctl status do a kill() test to verify that the PID found in
postmaster.pid still represents a live postmaster.
2005-04-20 23:10:16 +00:00
Tom Lane
1c155c8dfb Add note clarifying that indexes that support ordered scans had better
allow clauseless scans.
2005-04-20 22:19:58 +00:00
Tom Lane
eb4f58ad40 Don't try to run clauseless index scans on index types that don't support
it.  Per report from Marinos Yannikos.
2005-04-20 21:48:04 +00:00
Tom Lane
a8ac7d8713 Fix mis-display of negative fractional seconds in interval values for
--enable-integer-datetimes case.  Per report from Oliver Siegmar.
2005-04-20 17:14:50 +00:00
Tom Lane
9d64632144 Minor performance improvement: avoid unnecessary creation/unioning of
bitmaps for multiple indexscans.  Instead just let each indexscan add
TIDs directly into the BitmapOr node's result bitmap.
2005-04-20 15:48:36 +00:00
Bruce Momjian
de4fbfadc5 Add:
> * Add tool to query pg_stat_* tables and report indexes that aren't needed
>   or tables that might need indexes
2005-04-20 02:48:11 +00:00