1
0
mirror of https://github.com/postgres/postgres.git synced 2025-05-15 19:15:29 +03:00
postgres/src/backend/executor/nodeUnique.c
Peter Eisentraut 5e1963fb76 Collations with nondeterministic comparison
This adds a flag "deterministic" to collations.  If that is false,
such a collation disables various optimizations that assume that
strings are equal only if they are byte-wise equal.  That then allows
use cases such as case-insensitive or accent-insensitive comparisons
or handling of strings with different Unicode normal forms.

This functionality is only supported with the ICU provider.  At least
glibc doesn't appear to have any locales that work in a
nondeterministic way, so it's not worth supporting this for the libc
provider.

The term "deterministic comparison" in this context is from Unicode
Technical Standard #10
(https://unicode.org/reports/tr10/#Deterministic_Comparison).

This patch makes changes in three areas:

- CREATE COLLATION DDL changes and system catalog changes to support
  this new flag.

- Many executor nodes and auxiliary code are extended to track
  collations.  Previously, this code would just throw away collation
  information, because the eventually-called user-defined functions
  didn't use it since they only cared about equality, which didn't
  need collation information.

- String data type functions that do equality comparisons and hashing
  are changed to take the (non-)deterministic flag into account.  For
  comparison, this just means skipping various shortcuts and tie
  breakers that use byte-wise comparison.  For hashing, we first need
  to convert the input string to a canonical "sort key" using the ICU
  analogue of strxfrm().

Reviewed-by: Daniel Verite <daniel@manitou-mail.org>
Reviewed-by: Peter Geoghegan <pg@bowt.ie>
Discussion: https://www.postgresql.org/message-id/flat/1ccc668f-4cbc-0bef-af67-450b47cdfee7@2ndquadrant.com
2019-03-22 12:12:43 +01:00

193 lines
5.1 KiB
C

/*-------------------------------------------------------------------------
*
* nodeUnique.c
* Routines to handle unique'ing of queries where appropriate
*
* Unique is a very simple node type that just filters out duplicate
* tuples from a stream of sorted tuples from its subplan. It's essentially
* a dumbed-down form of Group: the duplicate-removal functionality is
* identical. However, Unique doesn't do projection nor qual checking,
* so it's marginally more efficient for cases where neither is needed.
* (It's debatable whether the savings justifies carrying two plan node
* types, though.)
*
* Portions Copyright (c) 1996-2019, PostgreSQL Global Development Group
* Portions Copyright (c) 1994, Regents of the University of California
*
*
* IDENTIFICATION
* src/backend/executor/nodeUnique.c
*
*-------------------------------------------------------------------------
*/
/*
* INTERFACE ROUTINES
* ExecUnique - generate a unique'd temporary relation
* ExecInitUnique - initialize node and subnodes
* ExecEndUnique - shutdown node and subnodes
*
* NOTES
* Assumes tuples returned from subplan arrive in
* sorted order.
*/
#include "postgres.h"
#include "executor/executor.h"
#include "executor/nodeUnique.h"
#include "miscadmin.h"
#include "utils/memutils.h"
/* ----------------------------------------------------------------
* ExecUnique
* ----------------------------------------------------------------
*/
static TupleTableSlot * /* return: a tuple or NULL */
ExecUnique(PlanState *pstate)
{
UniqueState *node = castNode(UniqueState, pstate);
ExprContext *econtext = node->ps.ps_ExprContext;
TupleTableSlot *resultTupleSlot;
TupleTableSlot *slot;
PlanState *outerPlan;
CHECK_FOR_INTERRUPTS();
/*
* get information from the node
*/
outerPlan = outerPlanState(node);
resultTupleSlot = node->ps.ps_ResultTupleSlot;
/*
* now loop, returning only non-duplicate tuples. We assume that the
* tuples arrive in sorted order so we can detect duplicates easily. The
* first tuple of each group is returned.
*/
for (;;)
{
/*
* fetch a tuple from the outer subplan
*/
slot = ExecProcNode(outerPlan);
if (TupIsNull(slot))
{
/* end of subplan, so we're done */
ExecClearTuple(resultTupleSlot);
return NULL;
}
/*
* Always return the first tuple from the subplan.
*/
if (TupIsNull(resultTupleSlot))
break;
/*
* Else test if the new tuple and the previously returned tuple match.
* If so then we loop back and fetch another new tuple from the
* subplan.
*/
econtext->ecxt_innertuple = slot;
econtext->ecxt_outertuple = resultTupleSlot;
if (!ExecQualAndReset(node->eqfunction, econtext))
break;
}
/*
* We have a new tuple different from the previous saved tuple (if any).
* Save it and return it. We must copy it because the source subplan
* won't guarantee that this source tuple is still accessible after
* fetching the next source tuple.
*/
return ExecCopySlot(resultTupleSlot, slot);
}
/* ----------------------------------------------------------------
* ExecInitUnique
*
* This initializes the unique node state structures and
* the node's subplan.
* ----------------------------------------------------------------
*/
UniqueState *
ExecInitUnique(Unique *node, EState *estate, int eflags)
{
UniqueState *uniquestate;
/* check for unsupported flags */
Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
/*
* create state structure
*/
uniquestate = makeNode(UniqueState);
uniquestate->ps.plan = (Plan *) node;
uniquestate->ps.state = estate;
uniquestate->ps.ExecProcNode = ExecUnique;
/*
* create expression context
*/
ExecAssignExprContext(estate, &uniquestate->ps);
/*
* then initialize outer plan
*/
outerPlanState(uniquestate) = ExecInitNode(outerPlan(node), estate, eflags);
/*
* Initialize result slot and type. Unique nodes do no projections, so
* initialize projection info for this node appropriately.
*/
ExecInitResultTupleSlotTL(&uniquestate->ps, &TTSOpsMinimalTuple);
uniquestate->ps.ps_ProjInfo = NULL;
/*
* Precompute fmgr lookup data for inner loop
*/
uniquestate->eqfunction =
execTuplesMatchPrepare(ExecGetResultType(outerPlanState(uniquestate)),
node->numCols,
node->uniqColIdx,
node->uniqOperators,
node->uniqCollations,
&uniquestate->ps);
return uniquestate;
}
/* ----------------------------------------------------------------
* ExecEndUnique
*
* This shuts down the subplan and frees resources allocated
* to this node.
* ----------------------------------------------------------------
*/
void
ExecEndUnique(UniqueState *node)
{
/* clean up tuple table */
ExecClearTuple(node->ps.ps_ResultTupleSlot);
ExecFreeExprContext(&node->ps);
ExecEndNode(outerPlanState(node));
}
void
ExecReScanUnique(UniqueState *node)
{
/* must clear result tuple so first input tuple is returned */
ExecClearTuple(node->ps.ps_ResultTupleSlot);
/*
* if chgParam of subnode is not null then plan will be re-scanned by
* first ExecProcNode.
*/
if (node->ps.lefttree->chgParam == NULL)
ExecReScan(node->ps.lefttree);
}