From bc9236bc014aea08e06d52516e47abed6fdcaa53 Mon Sep 17 00:00:00 2001
From: Tom Lane <tgl@sss.pgh.pa.us>
Date: Fri, 16 Jul 1999 22:17:06 +0000
Subject: [PATCH] Revise _bt_binsrch() so that its binary search loop takes
care of equal-key cases, eliminating bt_firsteq(). The linear search
formerly done by bt_firsteq() took a lot of time in the case where many equal
keys appear on the same page.
---
src/backend/access/nbtree/nbtsearch.c | 208 +++++++++++---------------
1 file changed, 85 insertions(+), 123 deletions(-)
diff --git a/src/backend/access/nbtree/nbtsearch.c b/src/backend/access/nbtree/nbtsearch.c
index 74c8b173fe3..c8ab46c8bca 100644
--- a/src/backend/access/nbtree/nbtsearch.c
+++ b/src/backend/access/nbtree/nbtsearch.c
@@ -7,7 +7,7 @@
*
*
* IDENTIFICATION
- * $Header: /cvsroot/pgsql/src/backend/access/nbtree/nbtsearch.c,v 1.50 1999/07/16 04:58:30 momjian Exp $
+ * $Header: /cvsroot/pgsql/src/backend/access/nbtree/nbtsearch.c,v 1.51 1999/07/16 22:17:06 tgl Exp $
*
*-------------------------------------------------------------------------
*/
@@ -26,8 +26,6 @@
static BTStack _bt_searchr(Relation rel, int keysz, ScanKey scankey,
Buffer *bufP, BTStack stack_in);
-static OffsetNumber _bt_firsteq(Relation rel, TupleDesc itupdesc, Page page,
- Size keysz, ScanKey scankey, OffsetNumber offnum);
static int _bt_compare(Relation rel, TupleDesc itupdesc, Page page,
int keysz, ScanKey scankey, OffsetNumber offnum);
static bool
@@ -368,7 +366,9 @@ _bt_skeycmp(Relation rel,
* comparison for every key in the scankey. _bt_binsrch() returns
* the OffsetNumber of the first matching key on the page, or the
* OffsetNumber at which the matching key would appear if it were
- * on this page.
+ * on this page. (NOTE: in particular, this means it is possible to
+ * return a value 1 greater than the number of keys on the page, if
+ * the scankey is > all keys on the page.)
*
* By the time this procedure is called, we're sure we're looking
* at the right page -- don't need to walk right. _bt_binsrch() has
@@ -385,8 +385,8 @@ _bt_binsrch(Relation rel,
Page page;
BTPageOpaque opaque;
OffsetNumber low,
- mid,
high;
+ bool haveEq;
int natts = rel->rd_rel->relnatts;
int result;
@@ -395,148 +395,112 @@ _bt_binsrch(Relation rel,
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
/* by convention, item 1 on any non-rightmost page is the high key */
- low = mid = P_RIGHTMOST(opaque) ? P_HIKEY : P_FIRSTKEY;
+ low = P_RIGHTMOST(opaque) ? P_HIKEY : P_FIRSTKEY;
high = PageGetMaxOffsetNumber(page);
/*
- * Since for non-rightmost pages, the first item on the page is the
- * high key, there are two notions of emptiness. One is if nothing
- * appears on the page. The other is if nothing but the high key
- * does. The reason we test high <= low, rather than high == low, is
- * that after vacuuming there may be nothing *but* the high key on a
- * page. In that case, given the scheme above, low = 2 and high = 1.
+ * If there are no keys on the page, return the first available slot.
+ * Note this covers two cases: the page is really empty (no keys),
+ * or it contains only a high key. The latter case is possible after
+ * vacuuming.
*/
-
- if (PageIsEmpty(page))
+ if (high < low)
return low;
- if ((!P_RIGHTMOST(opaque) && high <= low))
- {
- if (high < low ||
- (srchtype == BT_DESCENT && !(opaque->btpo_flags & BTP_LEAF)))
- return low;
- /* It's insertion and high == low == 2 */
- result = _bt_compare(rel, itupdesc, page, keysz, scankey, low);
- if (result > 0)
- return OffsetNumberNext(low);
- return low;
- }
- while ((high - low) > 1)
+ /*
+ * Binary search to find the first key on the page >= scan key.
+ * Loop invariant: all slots before 'low' are < scan key, all slots
+ * at or after 'high' are >= scan key. Also, haveEq is true if the
+ * tuple at 'high' is == scan key.
+ * We can fall out when high == low.
+ */
+ high++; /* establish the loop invariant for high */
+ haveEq = false;
+
+ while (high > low)
{
- mid = low + ((high - low) / 2);
+ OffsetNumber mid = low + ((high - low) / 2);
+ /* We have low <= mid < high, so mid points at a real slot */
+
result = _bt_compare(rel, itupdesc, page, keysz, scankey, mid);
if (result > 0)
- low = mid;
- else if (result < 0)
- high = mid - 1;
+ low = mid + 1;
else
{
- mid = _bt_firsteq(rel, itupdesc, page, keysz, scankey, mid);
-
- /*
- * NOTE for multi-column indices: we may do scan using keys
- * not for all attrs. But we handle duplicates using all attrs
- * in _bt_insert/_bt_spool code. And so while searching on
- * internal pages having number of attrs > keysize we want to
- * point at the last item < the scankey, not at the first item
- * = the scankey (!!!), and let _bt_moveright decide later
- * whether to move right or not (see comments and example
- * there). Note also that INSERTions are not affected by this
- * code (natts == keysz). - vadim 04/15/97
- */
- if (natts == keysz || opaque->btpo_flags & BTP_LEAF)
- return mid;
- low = P_RIGHTMOST(opaque) ? P_HIKEY : P_FIRSTKEY;
- if (mid == low)
- return mid;
- return OffsetNumberPrev(mid);
+ high = mid;
+ haveEq = (result == 0);
}
}
- /*
- * We terminated because the endpoints got too close together. There
- * are two cases to take care of.
+ /*--------------------
+ * At this point we have high == low, but be careful: they could point
+ * past the last slot on the page. We also know that haveEq is true
+ * if and only if there is an equal key (in which case high&low point
+ * at the first equal key).
*
- * For non-insertion searches on internal pages, we want to point at the
- * last key <, or first key =, the scankey on the page. This
- * guarantees that we'll descend the tree correctly. (NOTE comments
- * above for multi-column indices).
- *
- * For all other cases, we want to point at the first key >= the scankey
- * on the page. This guarantees that scans and insertions will happen
- * correctly.
+ * On a leaf page, we always return the first key >= scan key
+ * (which could be the last slot + 1).
+ *--------------------
*/
- if (!(opaque->btpo_flags & BTP_LEAF) && srchtype == BT_DESCENT)
- { /* We want the last key <, or first key
- * ==, the scan key. */
- result = _bt_compare(rel, itupdesc, page, keysz, scankey, high);
+ if (opaque->btpo_flags & BTP_LEAF)
+ return low;
- if (result == 0)
- {
- mid = _bt_firsteq(rel, itupdesc, page, keysz, scankey, high);
+ /*--------------------
+ * On a non-leaf page, there are special cases:
+ *
+ * For an insertion (srchtype != BT_DESCENT and natts == keysz)
+ * always return first key >= scan key (which could be off the end).
+ *
+ * For a standard search (srchtype == BT_DESCENT and natts == keysz)
+ * return the first equal key if one exists, else the last lesser key
+ * if one exists, else the first slot on the page.
+ *
+ * For a partial-match search (srchtype == BT_DESCENT and natts < keysz)
+ * return the last lesser key if one exists, else the first slot.
+ *
+ * Old comments:
+ * For multi-column indices, we may scan using keys
+ * not for all attrs. But we handle duplicates using all attrs
+ * in _bt_insert/_bt_spool code. And so while searching on
+ * internal pages having number of attrs > keysize we want to
+ * point at the last item < the scankey, not at the first item
+ * = the scankey (!!!), and let _bt_moveright decide later
+ * whether to move right or not (see comments and example
+ * there). Note also that INSERTions are not affected by this
+ * code (since natts == keysz for inserts). - vadim 04/15/97
+ *--------------------
+ */
- /*
- * If natts > keysz we want last item < the scan key. See
- * comments above for multi-column indices.
- */
- if (natts == keysz)
- return mid;
- low = P_RIGHTMOST(opaque) ? P_HIKEY : P_FIRSTKEY;
- if (mid == low)
- return mid;
- return OffsetNumberPrev(mid);
- }
- else if (result > 0)
- return high;
- else
- return low;
+ if (haveEq)
+ {
+ /*
+ * There is an equal key. We return either the first equal key
+ * (which we just found), or the last lesser key.
+ *
+ * We need not check srchtype != BT_DESCENT here, since if that
+ * is true then natts == keysz by assumption.
+ */
+ if (natts == keysz)
+ return low; /* return first equal key */
}
else
-/* we want the first key >= the scan key */
{
- result = _bt_compare(rel, itupdesc, page, keysz, scankey, low);
- if (result <= 0)
- return low;
- else
- {
- if (low == high)
- return OffsetNumberNext(low);
-
- result = _bt_compare(rel, itupdesc, page, keysz, scankey, high);
- if (result <= 0)
- return high;
- else
- return OffsetNumberNext(high);
- }
+ /*
+ * There is no equal key. We return either the first greater key
+ * (which we just found), or the last lesser key.
+ */
+ if (srchtype != BT_DESCENT)
+ return low; /* return first greater key */
}
-}
-static OffsetNumber
-_bt_firsteq(Relation rel,
- TupleDesc itupdesc,
- Page page,
- Size keysz,
- ScanKey scankey,
- OffsetNumber offnum)
-{
- BTPageOpaque opaque;
- OffsetNumber limit;
- opaque = (BTPageOpaque) PageGetSpecialPointer(page);
+ if (low == (P_RIGHTMOST(opaque) ? P_HIKEY : P_FIRSTKEY))
+ return low; /* there is no prior item */
- /* skip the high key, if any */
- limit = P_RIGHTMOST(opaque) ? P_HIKEY : P_FIRSTKEY;
-
- /* walk backwards looking for the first key in the chain of duplicates */
- while (offnum > limit
- && _bt_compare(rel, itupdesc, page,
- keysz, scankey, OffsetNumberPrev(offnum)) == 0)
- offnum = OffsetNumberPrev(offnum);
-
- return offnum;
+ return OffsetNumberPrev(low);
}
/*
@@ -571,7 +535,6 @@ _bt_compare(Relation rel,
{
Datum datum;
BTItem btitem;
- ItemId itemid;
IndexTuple itup;
BTPageOpaque opaque;
ScanKey entry;
@@ -589,12 +552,11 @@ _bt_compare(Relation rel,
*/
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
+
if (!(opaque->btpo_flags & BTP_LEAF)
&& P_LEFTMOST(opaque)
&& offnum == P_HIKEY)
{
- itemid = PageGetItemId(page, offnum);
-
/*
* we just have to believe that this will only be called with
* offnum == P_HIKEY when P_HIKEY is the OffsetNumber of the first
@@ -621,7 +583,7 @@ _bt_compare(Relation rel,
* on the page is greater than anything.
*/
- if (_bt_skeycmp(rel, keysz, scankey, page, itemid,
+ if (_bt_skeycmp(rel, keysz, scankey, page, PageGetItemId(page, offnum),
BTEqualStrategyNumber))
return 0;
return 1;
--
GitLab