tuptoaster.c

PostgreSQL 8.1.4的源码 适用于Linux下的开源数据库系统

/*-------------------------------------------------------------------------
 *
 * tuptoaster.c
 *	  Support routines for external and compressed storage of
 *	  variable size attributes.
 *
 * Copyright (c) 2000-2005, PostgreSQL Global Development Group
 *
 *
 * IDENTIFICATION
 *	  $PostgreSQL: pgsql/src/backend/access/heap/tuptoaster.c,v 1.53.2.3 2006/01/17 17:33:20 tgl Exp $
 *
 *
 * INTERFACE ROUTINES
 *		toast_insert_or_update -
 *			Try to make a given tuple fit into one page by compressing
 *			or moving off attributes
 *
 *		toast_delete -
 *			Reclaim toast storage when a tuple is deleted
 *
 *		heap_tuple_untoast_attr -
 *			Fetch back a given value from the "secondary" relation
 *
 *-------------------------------------------------------------------------
 */

#include "postgres.h"

#include <unistd.h>
#include <fcntl.h>

#include "access/heapam.h"
#include "access/genam.h"
#include "access/tuptoaster.h"
#include "catalog/catalog.h"
#include "utils/rel.h"
#include "utils/builtins.h"
#include "utils/fmgroids.h"
#include "utils/pg_lzcompress.h"
#include "utils/typcache.h"


#undef TOAST_DEBUG

static void toast_delete_datum(Relation rel, Datum value);
static Datum toast_save_datum(Relation rel, Datum value);
static varattrib *toast_fetch_datum(varattrib *attr);
static varattrib *toast_fetch_datum_slice(varattrib *attr,
						int32 sliceoffset, int32 length);


/* ----------
 * heap_tuple_fetch_attr -
 *
 *	Public entry point to get back a toasted value
 *	external storage (possibly still in compressed format).
 * ----------
 */
varattrib *
heap_tuple_fetch_attr(varattrib *attr)
{
	varattrib  *result;

	if (VARATT_IS_EXTERNAL(attr))
	{
		/*
		 * This is an external stored plain value
		 */
		result = toast_fetch_datum(attr);
	}
	else
	{
		/*
		 * This is a plain value inside of the main tuple - why am I called?
		 */
		result = attr;
	}

	return result;
}


/* ----------
 * heap_tuple_untoast_attr -
 *
 *	Public entry point to get back a toasted value from compression
 *	or external storage.
 * ----------
 */
varattrib *
heap_tuple_untoast_attr(varattrib *attr)
{
	varattrib  *result;

	if (VARATT_IS_EXTERNAL(attr))
	{
		if (VARATT_IS_COMPRESSED(attr))
		{
			/* ----------
			 * This is an external stored compressed value
			 * Fetch it from the toast heap and decompress.
			 * ----------
			 */
			varattrib  *tmp;

			tmp = toast_fetch_datum(attr);
			result = (varattrib *) palloc(attr->va_content.va_external.va_rawsize
										  + VARHDRSZ);
			VARATT_SIZEP(result) = attr->va_content.va_external.va_rawsize
				+ VARHDRSZ;
			pglz_decompress((PGLZ_Header *) tmp, VARATT_DATA(result));

			pfree(tmp);
		}
		else
		{
			/*
			 * This is an external stored plain value
			 */
			result = toast_fetch_datum(attr);
		}
	}
	else if (VARATT_IS_COMPRESSED(attr))
	{
		/*
		 * This is a compressed value inside of the main tuple
		 */
		result = (varattrib *) palloc(attr->va_content.va_compressed.va_rawsize
									  + VARHDRSZ);
		VARATT_SIZEP(result) = attr->va_content.va_compressed.va_rawsize
			+ VARHDRSZ;
		pglz_decompress((PGLZ_Header *) attr, VARATT_DATA(result));
	}
	else

		/*
		 * This is a plain value inside of the main tuple - why am I called?
		 */
		return attr;

	return result;
}


/* ----------
 * heap_tuple_untoast_attr_slice -
 *
 *		Public entry point to get back part of a toasted value
 *		from compression or external storage.
 * ----------
 */
varattrib *
heap_tuple_untoast_attr_slice(varattrib *attr, int32 sliceoffset, int32 slicelength)
{
	varattrib  *preslice;
	varattrib  *result;
	int32		attrsize;

	if (VARATT_IS_COMPRESSED(attr))
	{
		varattrib  *tmp;

		if (VARATT_IS_EXTERNAL(attr))
			tmp = toast_fetch_datum(attr);
		else
		{
			tmp = attr;			/* compressed in main tuple */
		}

		preslice = (varattrib *) palloc(attr->va_content.va_external.va_rawsize
										+ VARHDRSZ);
		VARATT_SIZEP(preslice) = attr->va_content.va_external.va_rawsize + VARHDRSZ;
		pglz_decompress((PGLZ_Header *) tmp, VARATT_DATA(preslice));

		if (tmp != attr)
			pfree(tmp);
	}
	else
	{
		/* Plain value */
		if (VARATT_IS_EXTERNAL(attr))
		{
			/* fast path */
			return (toast_fetch_datum_slice(attr, sliceoffset, slicelength));
		}
		else
			preslice = attr;
	}

	/* slicing of datum for compressed cases and plain value */

	attrsize = VARSIZE(preslice) - VARHDRSZ;
	if (sliceoffset >= attrsize)
	{
		sliceoffset = 0;
		slicelength = 0;
	}

	if (((sliceoffset + slicelength) > attrsize) || slicelength < 0)
		slicelength = attrsize - sliceoffset;

	result = (varattrib *) palloc(slicelength + VARHDRSZ);
	VARATT_SIZEP(result) = slicelength + VARHDRSZ;

	memcpy(VARDATA(result), VARDATA(preslice) + sliceoffset, slicelength);

	if (preslice != attr)
		pfree(preslice);

	return result;
}


/* ----------
 * toast_raw_datum_size -
 *
 *	Return the raw (detoasted) size of a varlena datum
 * ----------
 */
Size
toast_raw_datum_size(Datum value)
{
	varattrib  *attr = (varattrib *) DatumGetPointer(value);
	Size		result;

	if (VARATT_IS_COMPRESSED(attr))
	{
		/*
		 * va_rawsize shows the original data size, whether the datum is
		 * external or not.
		 */
		result = attr->va_content.va_compressed.va_rawsize + VARHDRSZ;
	}
	else if (VARATT_IS_EXTERNAL(attr))
	{
		/*
		 * an uncompressed external attribute has rawsize including the header
		 * (not too consistent!)
		 */
		result = attr->va_content.va_external.va_rawsize;
	}
	else
	{
		/* plain untoasted datum */
		result = VARSIZE(attr);
	}
	return result;
}

/* ----------
 * toast_datum_size
 *
 *	Return the physical storage size (possibly compressed) of a varlena datum
 * ----------
 */
Size
toast_datum_size(Datum value)
{
	varattrib  *attr = (varattrib *) DatumGetPointer(value);
	Size		result;

	if (VARATT_IS_EXTERNAL(attr))
	{
		/*
		 * Attribute is stored externally - return the extsize whether
		 * compressed or not.  We do not count the size of the toast pointer
		 * ... should we?
		 */
		result = attr->va_content.va_external.va_extsize;
	}
	else
	{
		/*
		 * Attribute is stored inline either compressed or not, just calculate
		 * the size of the datum in either case.
		 */
		result = VARSIZE(attr);
	}
	return result;
}


/* ----------
 * toast_delete -
 *
 *	Cascaded delete toast-entries on DELETE
 * ----------
 */
void
toast_delete(Relation rel, HeapTuple oldtup)
{
	TupleDesc	tupleDesc;
	Form_pg_attribute *att;
	int			numAttrs;
	int			i;
	Datum		toast_values[MaxHeapAttributeNumber];
	bool		toast_isnull[MaxHeapAttributeNumber];

	/*
	 * Get the tuple descriptor and break down the tuple into fields.
	 *
	 * NOTE: it's debatable whether to use heap_deformtuple() here or just
	 * heap_getattr() only the varlena columns.  The latter could win if there
	 * are few varlena columns and many non-varlena ones. However,
	 * heap_deformtuple costs only O(N) while the heap_getattr way would cost
	 * O(N^2) if there are many varlena columns, so it seems better to err on
	 * the side of linear cost.  (We won't even be here unless there's at
	 * least one varlena column, by the way.)
	 */
	tupleDesc = rel->rd_att;
	att = tupleDesc->attrs;
	numAttrs = tupleDesc->natts;

	Assert(numAttrs <= MaxHeapAttributeNumber);
	heap_deform_tuple(oldtup, tupleDesc, toast_values, toast_isnull);

	/*
	 * Check for external stored attributes and delete them from the secondary
	 * relation.
	 */
	for (i = 0; i < numAttrs; i++)
	{
		if (att[i]->attlen == -1)
		{
			Datum		value = toast_values[i];

			if (!toast_isnull[i] && VARATT_IS_EXTERNAL(value))
				toast_delete_datum(rel, value);
		}
	}
}


/* ----------
 * toast_insert_or_update -
 *
 *	Delete no-longer-used toast-entries and create new ones to
 *	make the new tuple fit on INSERT or UPDATE
 *
 * Inputs:
 *	newtup: the candidate new tuple to be inserted
 *	oldtup: the old row version for UPDATE, or NULL for INSERT
 * Result:
 *	either newtup if no toasting is needed, or a palloc'd modified tuple
 *	that is what should actually get stored
 *
 * NOTE: neither newtup nor oldtup will be modified.  This is a change
 * from the pre-8.1 API of this routine.
 * ----------
 */
HeapTuple
toast_insert_or_update(Relation rel, HeapTuple newtup, HeapTuple oldtup)
{
	HeapTuple	result_tuple;
	TupleDesc	tupleDesc;
	Form_pg_attribute *att;
	int			numAttrs;
	int			i;

	bool		need_change = false;
	bool		need_free = false;
	bool		need_delold = false;
	bool		has_nulls = false;

	Size		maxDataLen;

	char		toast_action[MaxHeapAttributeNumber];
	bool		toast_isnull[MaxHeapAttributeNumber];
	bool		toast_oldisnull[MaxHeapAttributeNumber];
	Datum		toast_values[MaxHeapAttributeNumber];
	Datum		toast_oldvalues[MaxHeapAttributeNumber];
	int32		toast_sizes[MaxHeapAttributeNumber];
	bool		toast_free[MaxHeapAttributeNumber];
	bool		toast_delold[MaxHeapAttributeNumber];

	/*
	 * Get the tuple descriptor and break down the tuple(s) into fields.
	 */
	tupleDesc = rel->rd_att;
	att = tupleDesc->attrs;
	numAttrs = tupleDesc->natts;

	Assert(numAttrs <= MaxHeapAttributeNumber);
	heap_deform_tuple(newtup, tupleDesc, toast_values, toast_isnull);
	if (oldtup != NULL)
		heap_deform_tuple(oldtup, tupleDesc, toast_oldvalues, toast_oldisnull);

	/* ----------
	 * Then collect information about the values given
	 *
	 * NOTE: toast_action[i] can have these values:
	 *		' '		default handling
	 *		'p'		already processed --- don't touch it
	 *		'x'		incompressible, but OK to move off
	 *
	 * NOTE: toast_sizes[i] is only made valid for varlena attributes with
	 *		toast_action[i] different from 'p'.
	 * ----------
	 */
	memset(toast_action, ' ', numAttrs * sizeof(char));
	memset(toast_free, 0, numAttrs * sizeof(bool));
	memset(toast_delold, 0, numAttrs * sizeof(bool));

	for (i = 0; i < numAttrs; i++)
	{
		varattrib  *old_value;
		varattrib  *new_value;

		if (oldtup != NULL)
		{
			/*
			 * For UPDATE get the old and new values of this attribute
			 */
			old_value = (varattrib *) DatumGetPointer(toast_oldvalues[i]);
			new_value = (varattrib *) DatumGetPointer(toast_values[i]);

			/*
			 * If the old value is an external stored one, check if it has
			 * changed so we have to delete it later.
			 */
			if (att[i]->attlen == -1 && !toast_oldisnull[i] &&
				VARATT_IS_EXTERNAL(old_value))
			{
				if (toast_isnull[i] || !VARATT_IS_EXTERNAL(new_value) ||
					old_value->va_content.va_external.va_valueid !=
					new_value->va_content.va_external.va_valueid ||
					old_value->va_content.va_external.va_toastrelid !=
					new_value->va_content.va_external.va_toastrelid)
				{
					/*
					 * The old external stored value isn't needed any more
					 * after the update
					 */
					toast_delold[i] = true;
					need_delold = true;
				}
				else
				{
					/*
					 * This attribute isn't changed by this update so we reuse
					 * the original reference to the old value in the new
					 * tuple.
					 */
					toast_action[i] = 'p';
					toast_sizes[i] = VARATT_SIZE(toast_values[i]);
					continue;
				}
			}
		}
		else
		{
			/*
			 * For INSERT simply get the new value
			 */
			new_value = (varattrib *) DatumGetPointer(toast_values[i]);
		}

		/*
		 * Handle NULL attributes
		 */
		if (toast_isnull[i])
		{
			toast_action[i] = 'p';
			has_nulls = true;
			continue;
		}

		/*
		 * Now look at varlena attributes
		 */
		if (att[i]->attlen == -1)
		{
			/*
			 * If the table's attribute says PLAIN always, force it so.
			 */
			if (att[i]->attstorage == 'p')
				toast_action[i] = 'p';

			/*
			 * We took care of UPDATE above, so any external value we find
			 * still in the tuple must be someone else's we cannot reuse.
			 * Expand it to plain (and, probably, toast it again below).
			 */
			if (VARATT_IS_EXTERNAL(new_value))
			{
				new_value = heap_tuple_untoast_attr(new_value);
				toast_values[i] = PointerGetDatum(new_value);
				toast_free[i] = true;
				need_change = true;
				need_free = true;