【标题】: 回合https://gitee.com/opengauss/openGauss-server/pulls/4694/修复表增加字段后core的问题.

【实现内容】: 回合https://gitee.com/opengauss/openGauss-server/pulls/4694/修复表增加字段后core的问题
【根因分析】: 回合https://gitee.com/opengauss/openGauss-server/pulls/4694/修复表增加字段后core的问题
【实现方案】: 回合https://gitee.com/opengauss/openGauss-server/pulls/4694/修复表增加字段后core的问题
【关联需求或issue】: https://e.gitee.com/opengaussorg/dashboard?issue=I9C345

src/common/backend/parser/parse_coerce.cpp

@@ -524,7 +524,7 @@ Node* coerce_type(ParseState* pstate, Node* node, Oid inputTypeId, Oid targetTyp
 
         newcon->consttype = baseTypeId;
         newcon->consttypmod = inputTypeMod;
-        if (OidIsValid(GetCollationConnection()) &&
+        if (!DB_IS_CMPT(B_FORMAT) && OidIsValid(GetCollationConnection()) &&
             IsSupportCharsetType(baseTypeId)) {
             newcon->constcollid = GetCollationConnection();
         } else {

src/common/backend/parser/parse_collate.cpp

@@ -136,7 +136,7 @@ FORCE_INLINE static void get_valid_collation(Oid& collation, CollateStrength& st
             strength = context.strength;
             location = context.location;
             derivation = context.derivation;
-        } else if (ENABLE_MULTI_CHARSET &&
+        } else if (!DB_IS_CMPT(B_FORMAT) && ENABLE_MULTI_CHARSET &&
             context.strength == COLLATE_NONE &&
             (OidIsValid(GetCollationConnection()) || context.derivation == DERIVATION_SYSCONST)) {
             /* collation of version() is UTF8_GENERAL_CI */

src/common/backend/parser/parse_node.cpp

@@ -520,7 +520,7 @@ Const* make_const(ParseState* pstate, Value* value, int location)
             typid = UNKNOWNOID; /* will be coerced later */
             typelen = -2;       /* cstring-style varwidth type */
             typebyval = false;
-            if (OidIsValid(GetCollationConnection())) {
+            if (!DB_IS_CMPT(B_FORMAT) && OidIsValid(GetCollationConnection())) {
                 collid = GetCollationConnection();
             }
             break;

src/common/backend/utils/mb/mbutils.cpp

@@ -1218,7 +1218,7 @@ const char* GetCharsetConnectionName(void)
 
 Oid GetCollationConnection(void)
 {
-    if (!ENABLE_MULTI_CHARSET) {
+    if (!ENABLE_MULTI_CHARSET || !DB_IS_CMPT(B_FORMAT)) {
         return InvalidOid;
     }
     return u_sess->mb_cxt.collation_connection;

src/test/regress/expected/test_b_format_collate2.out

@@ -0,0 +1,114 @@
+CREATE DATABASE test_collate_A4 with dbcompatibility 'b' ENCODING='SQL_ASCII' LC_COLLATE='C' LC_CTYPE='C';
+\c test_collate_A4
+set b_format_behavior_compat_options = 'default_collation,enable_multi_charset';
+create or replace function pg_catalog.get_index_columns(OUT namespace name, OUT indexrelid oid, OUT indrelid oid, OUT indisunique bool, OUT indisusable bool, OUT seq_in_index int2, OUT attrnum int2, OUT collation int2) returns setof record
+as $$
+declare
+query_str text;
+item int2;
+row_data record;
+begin
+query_str := 'select n.nspname, i.indexrelid, i.indrelid, i.indisunique, i.indisusable, i.indkey, i.indoption, i.indnkeyatts
+              from pg_catalog.pg_index i
+                left join pg_class c on c.oid = i.indexrelid
+                left join pg_catalog.pg_namespace n on n.oid = c.relnamespace
+              where n.nspname <> ''pg_catalog''
+                and n.nspname <> ''db4ai''
+                and n.nspname <> ''information_schema''
+                and n.nspname !~ ''^pg_toast''';
+for row_data in EXECUTE(query_str) LOOP
+    for item in 0..row_data.indnkeyatts - 1 loop
+        namespace := row_data.nspname;
+        indexrelid := row_data.indexrelid;
+        indrelid := row_data.indrelid;
+        indisunique := row_data.indisunique;
+        indisusable := row_data.indisusable;
+        seq_in_index := item + 1;
+        attrnum := row_data.indkey[item];
+        collation := row_data.indoption[item];
+        return next;
+    end loop;
+end loop;
+end; $$
+LANGUAGE 'plpgsql';
+create view public.index_statistic as
+select
+  i.namespace as "namespace",
+  (select relname from pg_class tc where tc.oid = i.indrelid) as "table",
+  not i.indisunique as "non_unique",
+  c.relname as "key_name",
+  i.seq_in_index as "seq_in_index",
+  a.attname as "column_name",
+  (case when m.amcanorder
+    then (
+      case when i.collation & 1 then 'D' else 'A' END
+    ) else null end
+  ) as "collation",
+  (select
+      (case when ts.stadistinct = 0
+        then NULL else (
+          case when ts.stadistinct > 0 then ts.stadistinct else ts.stadistinct * tc.reltuples * -1 end
+        ) end
+      )
+    from pg_class tc
+      left join pg_statistic ts on tc.oid = ts.starelid
+    where
+      tc.oid = i.indrelid
+      and ts.staattnum = i.attrnum
+  ) as "cardinality",
+  null as "sub_part",
+  null as "packed",
+  (case when a.attnotnull then '' else 'YES' end) as "null",
+  m.amname as "index_type",
+  (case when i.indisusable then '' else 'disabled' end) as "comment",
+  (select description from pg_description where objoid = i.indexrelid) as "index_comment"
+from
+  (select * from get_index_columns()) i
+  left join pg_class c on c.oid = i.indexrelid
+  left join pg_attribute a on a.attrelid = i.indrelid
+  and a.attnum = i.attrnum
+  left join pg_am m on m.oid = c.relam
+order by
+  c.relname;
+CREATE OR REPLACE FUNCTION pg_catalog.pg_get_nonstrict_basic_value(typename text)
+RETURNS text
+AS
+$$
+BEGIN
+    IF typename = 'timestamp' then
+        return 'now';
+    elsif typename = 'time' or typename = 'timetz' or typename = 'interval' or typename = 'reltime' then
+        return '00:00:00';
+    elsif typename = 'date' then
+        return '1970-01-01';
+    elsif typename = 'smalldatetime' then
+        return '1970-01-01 08:00:00';
+    elsif typename = 'abstime' then
+        return '1970-01-01 08:00:00+08';
+    elsif typename = 'uuid' then
+        return '00000000-0000-0000-0000-000000000000';
+    elsif typename = 'bool' then
+        return 'false';
+    elsif typename = 'point' or typename = 'polygon' then
+        return '(0,0)';
+    elsif typename = 'path' then
+        return '((0,0))';
+    elsif typename = 'circle' then
+        return '(0,0),0';
+    elsif typename = 'lseg' or typename = 'box' then
+        return '(0,0),(0,0)';
+    elsif typename = 'tinterval' then
+        return '["1970-01-01 00:00:00+08" "1970-01-01 00:00:00+08"]';
+    else
+        return '0 or empty';
+    end if;
+end;
+$$
+LANGUAGE plpgsql;
+CREATE VIEW public.pg_type_nonstrict_basic_value AS
+    SELECT
+            t.typname As typename,
+            pg_get_nonstrict_basic_value(t.typname) As basic_value
+    FROM pg_type t;
+\c postgres
+DROP DATABASE IF EXISTS test_collate_A4;

src/test/regress/parallel_schedule0A

@@ -98,7 +98,7 @@ test: event_dump_audit
 test: single_node_sha
 
 #test b format collation
-test: test_b_format_collate charset_b_format
+test: test_b_format_collate charset_b_format test_b_format_collate2
 
 # test subpartition
 test: hw_subpartition_createtable hw_subpartition_scan hw_subpartition_select hw_subpartition_split hw_subpartition_truncate hw_subpartition_update hw_subpartition_gpi hw_subpartition_analyze_vacuum hw_subpartition_alter_table hw_subpartition_index hw_subpartition_add_drop_partition hw_subpartition_tablespace hw_subpartition_ddl_index hw_subpartition_size

src/test/regress/sql/test_b_format_collate2.sql

@@ -0,0 +1,123 @@
+CREATE DATABASE test_collate_A4 with dbcompatibility 'b' ENCODING='SQL_ASCII' LC_COLLATE='C' LC_CTYPE='C';
+\c test_collate_A4
+
+set b_format_behavior_compat_options = 'default_collation,enable_multi_charset';
+
+create or replace function pg_catalog.get_index_columns(OUT namespace name, OUT indexrelid oid, OUT indrelid oid, OUT indisunique bool, OUT indisusable bool, OUT seq_in_index int2, OUT attrnum int2, OUT collation int2) returns setof record
+as $$
+declare
+query_str text;
+item int2;
+row_data record;
+begin
+query_str := 'select n.nspname, i.indexrelid, i.indrelid, i.indisunique, i.indisusable, i.indkey, i.indoption, i.indnkeyatts
+              from pg_catalog.pg_index i
+                left join pg_class c on c.oid = i.indexrelid
+                left join pg_catalog.pg_namespace n on n.oid = c.relnamespace
+              where n.nspname <> ''pg_catalog''
+                and n.nspname <> ''db4ai''
+                and n.nspname <> ''information_schema''
+                and n.nspname !~ ''^pg_toast''';
+for row_data in EXECUTE(query_str) LOOP
+    for item in 0..row_data.indnkeyatts - 1 loop
+        namespace := row_data.nspname;
+        indexrelid := row_data.indexrelid;
+        indrelid := row_data.indrelid;
+        indisunique := row_data.indisunique;
+        indisusable := row_data.indisusable;
+        seq_in_index := item + 1;
+        attrnum := row_data.indkey[item];
+        collation := row_data.indoption[item];
+        return next;
+    end loop;
+end loop;
+end; $$
+LANGUAGE 'plpgsql';
+
+create view public.index_statistic as
+select
+  i.namespace as "namespace",
+  (select relname from pg_class tc where tc.oid = i.indrelid) as "table",
+  not i.indisunique as "non_unique",
+  c.relname as "key_name",
+  i.seq_in_index as "seq_in_index",
+  a.attname as "column_name",
+  (case when m.amcanorder
+    then (
+      case when i.collation & 1 then 'D' else 'A' END
+    ) else null end
+  ) as "collation",
+  (select
+      (case when ts.stadistinct = 0
+        then NULL else (
+          case when ts.stadistinct > 0 then ts.stadistinct else ts.stadistinct * tc.reltuples * -1 end
+        ) end
+      )
+    from pg_class tc
+      left join pg_statistic ts on tc.oid = ts.starelid
+    where
+      tc.oid = i.indrelid
+      and ts.staattnum = i.attrnum
+  ) as "cardinality",
+  null as "sub_part",
+  null as "packed",
+  (case when a.attnotnull then '' else 'YES' end) as "null",
+  m.amname as "index_type",
+  (case when i.indisusable then '' else 'disabled' end) as "comment",
+  (select description from pg_description where objoid = i.indexrelid) as "index_comment"
+from
+  (select * from get_index_columns()) i
+  left join pg_class c on c.oid = i.indexrelid
+  left join pg_attribute a on a.attrelid = i.indrelid
+  and a.attnum = i.attrnum
+  left join pg_am m on m.oid = c.relam
+order by
+  c.relname;
+
+
+CREATE OR REPLACE FUNCTION pg_catalog.pg_get_nonstrict_basic_value(typename text)
+RETURNS text
+AS
+$$
+BEGIN
+    IF typename = 'timestamp' then
+        return 'now';
+    elsif typename = 'time' or typename = 'timetz' or typename = 'interval' or typename = 'reltime' then
+        return '00:00:00';
+    elsif typename = 'date' then
+        return '1970-01-01';
+    elsif typename = 'smalldatetime' then
+        return '1970-01-01 08:00:00';
+    elsif typename = 'abstime' then
+        return '1970-01-01 08:00:00+08';
+    elsif typename = 'uuid' then
+        return '00000000-0000-0000-0000-000000000000';
+    elsif typename = 'bool' then
+        return 'false';
+    elsif typename = 'point' or typename = 'polygon' then
+        return '(0,0)';
+    elsif typename = 'path' then
+        return '((0,0))';
+    elsif typename = 'circle' then
+        return '(0,0),0';
+    elsif typename = 'lseg' or typename = 'box' then
+        return '(0,0),(0,0)';
+    elsif typename = 'tinterval' then
+        return '["1970-01-01 00:00:00+08" "1970-01-01 00:00:00+08"]';
+    else
+        return '0 or empty';
+    end if;
+end;
+$$
+LANGUAGE plpgsql;
+
+CREATE VIEW public.pg_type_nonstrict_basic_value AS
+    SELECT
+            t.typname As typename,
+            pg_get_nonstrict_basic_value(t.typname) As basic_value
+
+    FROM pg_type t;
+
+
+\c postgres
+DROP DATABASE IF EXISTS test_collate_A4;