Skip to content

Commit 05649b8

Browse files
tglsfdctglsfdc
committed
Rewrite pg_dump's comment-dumping code to pull over all the comments
in one query, rather than making a separate query for each object that could have a comment. This costs relatively little space (a few tens of K typically) and saves substantial time in databases with many objects. I find it reduces the runtime of 'pg_dump -s regression' by about a third.
1 parent c8aaa5c commit 05649b8

File tree

1 file changed

+215
-93
lines changed

1 file changed

+215
-93
lines changed

‎src/bin/pg_dump/pg_dump.c‎

Lines changed: 215 additions & 93 deletions
Original file line numberDiff line numberDiff line change
@@ -12,7 +12,7 @@
1212
* by PostgreSQL
1313
*
1414
* IDENTIFICATION
15-
* $PostgreSQL: pgsql/src/bin/pg_dump/pg_dump.c,v 1.367 2004/03/03 21:28:54 tgl Exp $
15+
* $PostgreSQL: pgsql/src/bin/pg_dump/pg_dump.c,v 1.368 2004/03/20 20:09:45 tgl Exp $
1616
*
1717
*-------------------------------------------------------------------------
1818
*/
@@ -67,6 +67,15 @@ extern int optind,
6767
opterr;
6868

6969

70+
typedef struct
71+
{
72+
const char *descr; /* comment for an object */
73+
Oid classoid; /* object class (catalog OID) */
74+
Oid objoid; /* object OID */
75+
int objsubid; /* subobject (table column #) */
76+
} CommentItem;
77+
78+
7079
/* global decls */
7180
bool g_verbose; /* User wants verbose narration of our
7281
* activities. */
@@ -105,6 +114,9 @@ static void dumpTableData(Archive *fout, TableDataInfo *tdinfo);
105114
static void dumpComment(Archive *fout, const char *target,
106115
const char *namespace, const char *owner,
107116
CatalogId catalogId, int subid, DumpId dumpId);
117+
static int findComments(Archive *fout, Oid classoid, Oid objoid,
118+
CommentItem **items);
119+
static int collectComments(Archive *fout, CommentItem **items);
108120
static void dumpDumpableObject(Archive *fout, DumpableObject *dobj);
109121
static void dumpNamespace(Archive *fout, NamespaceInfo *nspinfo);
110122
static void dumpType(Archive *fout, TypeInfo *tinfo);
@@ -3880,141 +3892,81 @@ dumpComment(Archive *fout, const char *target,
38803892
const char *namespace, const char *owner,
38813893
CatalogId catalogId, int subid, DumpId dumpId)
38823894
{
3883-
PGresult *res;
3884-
PQExpBuffer query;
3885-
int i_description;
3895+
CommentItem *comments;
3896+
int ncomments;
38863897

38873898
/* Comments are SCHEMA not data */
38883899
if (dataOnly)
38893900
return;
38903901

3891-
/*
3892-
* Note we do NOT change source schema here; preserve the caller's
3893-
* setting, instead.
3894-
*/
3902+
/* Search for comments associated with catalogId, using table */
3903+
ncomments = findComments(fout, catalogId.tableoid, catalogId.oid,
3904+
&comments);
38953905

3896-
/* Build query to find comment */
3897-
3898-
query = createPQExpBuffer();
3899-
3900-
if (fout->remoteVersion >= 70300)
3901-
{
3902-
appendPQExpBuffer(query,
3903-
"SELECT description FROM pg_catalog.pg_description "
3904-
"WHERE classoid = '%u'::pg_catalog.oid and "
3905-
"objoid = '%u'::pg_catalog.oid and objsubid = %d",
3906-
catalogId.tableoid, catalogId.oid, subid);
3907-
}
3908-
else if (fout->remoteVersion >= 70200)
3909-
{
3910-
appendPQExpBuffer(query,
3911-
"SELECT description FROM pg_description "
3912-
"WHERE classoid = '%u'::oid and "
3913-
"objoid = '%u'::oid and objsubid = %d",
3914-
catalogId.tableoid, catalogId.oid, subid);
3915-
}
3916-
else
3906+
/* Is there one matching the subid? */
3907+
while (ncomments > 0)
39173908
{
3918-
/* Note: this will fail to find attribute comments in pre-7.2... */
3919-
appendPQExpBuffer(query, "SELECT description FROM pg_description WHERE objoid = '%u'::oid", catalogId.oid);
3909+
if (comments->objsubid == subid)
3910+
break;
3911+
comments++;
3912+
ncomments--;
39203913
}
39213914

3922-
/* Execute query */
3923-
3924-
res = PQexec(g_conn, query->data);
3925-
check_sql_result(res, g_conn, query->data, PGRES_TUPLES_OK);
3926-
39273915
/* If a comment exists, build COMMENT ON statement */
3928-
3929-
if (PQntuples(res) == 1)
3916+
if (ncomments > 0)
39303917
{
3931-
i_description = PQfnumber(res, "description");
3932-
resetPQExpBuffer(query);
3918+
PQExpBuffer query = createPQExpBuffer();
3919+
39333920
appendPQExpBuffer(query, "COMMENT ON %s IS ", target);
3934-
appendStringLiteral(query, PQgetvalue(res, 0, i_description), false);
3921+
appendStringLiteral(query, comments->descr, false);
39353922
appendPQExpBuffer(query, ";\n");
39363923

39373924
ArchiveEntry(fout, nilCatalogId, createDumpId(),
39383925
target, namespace, owner,
39393926
"COMMENT", query->data, "", NULL,
39403927
&(dumpId), 1,
39413928
NULL, NULL);
3942-
}
39433929

3944-
PQclear(res);
3945-
destroyPQExpBuffer(query);
3930+
destroyPQExpBuffer(query);
3931+
}
39463932
}
39473933

39483934
/*
39493935
* dumpTableComment --
39503936
*
39513937
* As above, but dump comments for both the specified table (or view)
3952-
* and its columns. For speed, we want to do this with only one query.
3938+
* and its columns.
39533939
*/
39543940
static void
39553941
dumpTableComment(Archive *fout, TableInfo *tbinfo,
39563942
const char *reltypename)
39573943
{
3958-
PGresult *res;
3944+
CommentItem *comments;
3945+
int ncomments;
39593946
PQExpBuffer query;
39603947
PQExpBuffer target;
3961-
int i_description;
3962-
int i_objsubid;
3963-
int ntups;
3964-
int i;
39653948

39663949
/* Comments are SCHEMA not data */
39673950
if (dataOnly)
39683951
return;
39693952

3970-
/*
3971-
* Note we do NOT change source schema here; preserve the caller's
3972-
* setting, instead.
3973-
*/
3953+
/* Search for comments associated with relation, using table */
3954+
ncomments = findComments(fout,
3955+
tbinfo->dobj.catId.tableoid,
3956+
tbinfo->dobj.catId.oid,
3957+
&comments);
39743958

3975-
/* Build query to find comments */
3959+
/* If comments exist, build COMMENT ON statements */
3960+
if (ncomments <= 0)
3961+
return;
39763962

39773963
query = createPQExpBuffer();
39783964
target = createPQExpBuffer();
39793965

3980-
if (fout->remoteVersion >= 70300)
3981-
{
3982-
appendPQExpBuffer(query, "SELECT description, objsubid FROM pg_catalog.pg_description "
3983-
"WHERE classoid = '%u'::pg_catalog.oid and "
3984-
"objoid = '%u'::pg_catalog.oid "
3985-
"ORDER BY objoid, classoid, objsubid",
3986-
tbinfo->dobj.catId.tableoid, tbinfo->dobj.catId.oid);
3987-
}
3988-
else if (fout->remoteVersion >= 70200)
3989-
{
3990-
appendPQExpBuffer(query, "SELECT description, objsubid FROM pg_description "
3991-
"WHERE classoid = '%u'::oid and "
3992-
"objoid = '%u'::oid "
3993-
"ORDER BY objoid, classoid, objsubid",
3994-
tbinfo->dobj.catId.tableoid, tbinfo->dobj.catId.oid);
3995-
}
3996-
else
3997-
{
3998-
/* Note: this will fail to find attribute comments in pre-7.2... */
3999-
appendPQExpBuffer(query, "SELECT description, 0 as objsubid FROM pg_description WHERE objoid = '%u'::oid",
4000-
tbinfo->dobj.catId.oid);
4001-
}
4002-
4003-
/* Execute query */
4004-
4005-
res = PQexec(g_conn, query->data);
4006-
check_sql_result(res, g_conn, query->data, PGRES_TUPLES_OK);
4007-
4008-
i_description = PQfnumber(res, "description");
4009-
i_objsubid = PQfnumber(res, "objsubid");
4010-
4011-
/* If comments exist, build COMMENT ON statements */
4012-
4013-
ntups = PQntuples(res);
4014-
for (i = 0; i < ntups; i++)
3966+
while (ncomments > 0)
40153967
{
4016-
const char *descr = PQgetvalue(res, i, i_description);
4017-
int objsubid = atoi(PQgetvalue(res, i, i_objsubid));
3968+
const char *descr = comments->descr;
3969+
int objsubid = comments->objsubid;
40183970

40193971
if (objsubid == 0)
40203972
{
@@ -4054,13 +4006,183 @@ dumpTableComment(Archive *fout, TableInfo *tbinfo,
40544006
&(tbinfo->dobj.dumpId), 1,
40554007
NULL, NULL);
40564008
}
4009+
4010+
comments++;
4011+
ncomments--;
40574012
}
40584013

4059-
PQclear(res);
40604014
destroyPQExpBuffer(query);
40614015
destroyPQExpBuffer(target);
40624016
}
40634017

4018+
/*
4019+
* findComments --
4020+
*
4021+
* Find the comment(s), if any, associated with the given object. All the
4022+
* objsubid values associated with the given classoid/objoid are found with
4023+
* one search.
4024+
*/
4025+
static int
4026+
findComments(Archive *fout, Oid classoid, Oid objoid,
4027+
CommentItem **items)
4028+
{
4029+
/* static storage for table of comments */
4030+
static CommentItem *comments = NULL;
4031+
static int ncomments = -1;
4032+
4033+
CommentItem *middle = NULL;
4034+
CommentItem *low;
4035+
CommentItem *high;
4036+
int nmatch;
4037+
4038+
/* Get comments if we didn't already */
4039+
if (ncomments < 0)
4040+
ncomments = collectComments(fout, &comments);
4041+
4042+
/*
4043+
* Pre-7.2, pg_description does not contain classoid, so collectComments
4044+
* just stores a zero. If there's a collision on object OID, well, you
4045+
* get duplicate comments.
4046+
*/
4047+
if (fout->remoteVersion < 70200)
4048+
classoid = 0;
4049+
4050+
/*
4051+
* Do binary search to find some item matching the object.
4052+
*/
4053+
low = &comments[0];
4054+
high = &comments[ncomments-1];
4055+
while (low <= high)
4056+
{
4057+
middle = low + (high - low) / 2;
4058+
4059+
if (classoid < middle->classoid)
4060+
high = middle - 1;
4061+
else if (classoid > middle->classoid)
4062+
low = middle + 1;
4063+
else if (objoid < middle->objoid)
4064+
high = middle - 1;
4065+
else if (objoid > middle->objoid)
4066+
low = middle + 1;
4067+
else
4068+
break; /* found a match */
4069+
}
4070+
4071+
if (low > high) /* no matches */
4072+
{
4073+
*items = NULL;
4074+
return 0;
4075+
}
4076+
4077+
/*
4078+
* Now determine how many items match the object. The search loop
4079+
* invariant still holds: only items between low and high inclusive
4080+
* could match.
4081+
*/
4082+
nmatch = 1;
4083+
while (middle > low)
4084+
{
4085+
if (classoid != middle[-1].classoid ||
4086+
objoid != middle[-1].objoid)
4087+
break;
4088+
middle--;
4089+
nmatch++;
4090+
}
4091+
4092+
*items = middle;
4093+
4094+
middle += nmatch;
4095+
while (middle <= high)
4096+
{
4097+
if (classoid != middle->classoid ||
4098+
objoid != middle->objoid)
4099+
break;
4100+
middle++;
4101+
nmatch++;
4102+
}
4103+
4104+
return nmatch;
4105+
}
4106+
4107+
/*
4108+
* collectComments --
4109+
*
4110+
* Construct a table of all comments available for database objects.
4111+
* We used to do per-object queries for the comments, but it's much faster
4112+
* to pull them all over at once, and on most databases the memory cost
4113+
* isn't high.
4114+
*
4115+
* The table is sorted by classoid/objid/objsubid for speed in lookup.
4116+
*/
4117+
static int
4118+
collectComments(Archive *fout, CommentItem **items)
4119+
{
4120+
PGresult *res;
4121+
PQExpBuffer query;
4122+
int i_description;
4123+
int i_classoid;
4124+
int i_objoid;
4125+
int i_objsubid;
4126+
int ntups;
4127+
int i;
4128+
CommentItem *comments;
4129+
4130+
/*
4131+
* Note we do NOT change source schema here; preserve the caller's
4132+
* setting, instead.
4133+
*/
4134+
4135+
query = createPQExpBuffer();
4136+
4137+
if (fout->remoteVersion >= 70300)
4138+
{
4139+
appendPQExpBuffer(query, "SELECT description, classoid, objoid, objsubid "
4140+
"FROM pg_catalog.pg_description "
4141+
"ORDER BY classoid, objoid, objsubid");
4142+
}
4143+
else if (fout->remoteVersion >= 70200)
4144+
{
4145+
appendPQExpBuffer(query, "SELECT description, classoid, objoid, objsubid "
4146+
"FROM pg_description "
4147+
"ORDER BY classoid, objoid, objsubid");
4148+
}
4149+
else
4150+
{
4151+
/* Note: this will fail to find attribute comments in pre-7.2... */
4152+
appendPQExpBuffer(query, "SELECT description, 0 as classoid, objoid, 0 as objsubid "
4153+
"FROM pg_description "
4154+
"ORDER BY objoid");
4155+
}
4156+
4157+
res = PQexec(g_conn, query->data);
4158+
check_sql_result(res, g_conn, query->data, PGRES_TUPLES_OK);
4159+
4160+
/* Construct lookup table containing OIDs in numeric form */
4161+
4162+
i_description = PQfnumber(res, "description");
4163+
i_classoid = PQfnumber(res, "classoid");
4164+
i_objoid = PQfnumber(res, "objoid");
4165+
i_objsubid = PQfnumber(res, "objsubid");
4166+
4167+
ntups = PQntuples(res);
4168+
4169+
comments = (CommentItem *) malloc(ntups * sizeof(CommentItem));
4170+
4171+
for (i = 0; i < ntups; i++)
4172+
{
4173+
comments[i].descr = PQgetvalue(res, i, i_description);
4174+
comments[i].classoid = atooid(PQgetvalue(res, i, i_classoid));
4175+
comments[i].objoid = atooid(PQgetvalue(res, i, i_objoid));
4176+
comments[i].objsubid = atoi(PQgetvalue(res, i, i_objsubid));
4177+
}
4178+
4179+
/* Do NOT free the PGresult since we are keeping pointers into it */
4180+
destroyPQExpBuffer(query);
4181+
4182+
*items = comments;
4183+
return ntups;
4184+
}
4185+
40644186
/*
40654187
* dumpDumpableObject
40664188
*

0 commit comments

Comments
 (0)