ii regexp: support multiple matches case

groonga · Apr 12, 2017 · fc89271 · fc89271
1 parent 96862f6
commit fc89271
Show file tree

Hide file tree

Showing 3 changed files with 93 additions and 2 deletions.
diff --git a/lib/ii.c b/lib/ii.c
@@ -7974,8 +7974,10 @@ grn_ii_select_cursor_next(grn_ctx *ctx,
         cursor->posting.pos = pos;
         cursor->posting.tf = n_occurs;
         cursor->posting.weight = tscore;
-        if (token_info_skip(ctx, *tis, next_rid, next_sid) != GRN_SUCCESS) {
-          cursor->done = GRN_TRUE;
+        if (token_info_skip_pos(ctx, *tis, rid, sid, pos + 1) != GRN_SUCCESS) {
+          if (token_info_skip(ctx, *tis, next_rid, next_sid) != GRN_SUCCESS) {
+            cursor->done = GRN_TRUE;
+          }
         }
         return &(cursor->posting);
       }
@@ -8168,6 +8170,8 @@ grn_ii_select_regexp(grn_ctx *ctx, grn_ii *ii,
           if (posting_i->rid == posting->rid &&
               posting_i->sid == posting->sid &&
               posting_i->pos > pos) {
+            keep_i = i;
+            keep_posting = *posting_i;
             break;
           }
           if (posting_i->rid > posting->rid) {

diff --git a/test/command/suite/select/filter/index/regexp/dot_asterisk/n_matches.expected b/test/command/suite/select/filter/index/regexp/dot_asterisk/n_matches.expected
@@ -0,0 +1,62 @@
+table_create Memos TABLE_NO_KEY
+[[0,0.0,0.0],true]
+column_create Memos content COLUMN_SCALAR Text
+[[0,0.0,0.0],true]
+table_create RegexpTokens TABLE_PAT_KEY ShortText   --normalizer NormalizerAuto   --default_tokenizer TokenRegexp
+[[0,0.0,0.0],true]
+column_create RegexpTokens memos_content COLUMN_INDEX|WITH_POSITION   Memos content
+[[0,0.0,0.0],true]
+load --table Memos
+[
+{"content": "Groonga Groonga"},
+{"content": "Groonga Rroonga"},
+{"content": "PGroonga PostgreSQL"}
+]
+[[0,0.0,0.0],3]
+log_level --level info
+[[0,0.0,0.0],true]
+select Memos   --filter 'content @~ "g.*ga"'   --output_columns content,_score
+[
+  [
+    0,
+    0.0,
+    0.0
+  ],
+  [
+    [
+      [
+        3
+      ],
+      [
+        [
+          "content",
+          "Text"
+        ],
+        [
+          "_score",
+          "Int32"
+        ]
+      ],
+      [
+        "Groonga Groonga",
+        3
+      ],
+      [
+        "Groonga Rroonga",
+        2
+      ],
+      [
+        "PGroonga PostgreSQL",
+        1
+      ]
+    ]
+  ]
+]
+#|i| [object][search][index][key][regexp] <RegexpTokens.memos_content>
+#|i| grn_ii_sel > (g.*ga)
+#|i| [ii][select][cursor][open] n=1 <g>
+#|i| [ii][select][cursor][open] n=1 <ga>
+#|i| exact: 3
+#|i| hits=3
+log_level --level notice
+[[0,0.0,0.0],true]
diff --git a/test/command/suite/select/filter/index/regexp/dot_asterisk/n_matches.test b/test/command/suite/select/filter/index/regexp/dot_asterisk/n_matches.test
@@ -0,0 +1,25 @@
+#$GRN_II_REGEXP_DOT_ASTERISK_ENABLE=yes
+
+table_create Memos TABLE_NO_KEY
+column_create Memos content COLUMN_SCALAR Text
+
+table_create RegexpTokens TABLE_PAT_KEY ShortText \
+  --normalizer NormalizerAuto \
+  --default_tokenizer TokenRegexp
+column_create RegexpTokens memos_content COLUMN_INDEX|WITH_POSITION \
+  Memos content
+
+load --table Memos
+[
+{"content": "Groonga Groonga"},
+{"content": "Groonga Rroonga"},
+{"content": "PGroonga PostgreSQL"}
+]
+
+log_level --level info
+#@add-important-log-levels info
+select Memos \
+  --filter 'content @~ "g.*ga"' \
+  --output_columns content,_score
+#@remove-important-log-levels info
+log_level --level notice