summaryrefslogtreecommitdiff
path: root/storage/innobase/fts
diff options
context:
space:
mode:
Diffstat (limited to 'storage/innobase/fts')
-rw-r--r--storage/innobase/fts/fts0ast.cc34
-rw-r--r--storage/innobase/fts/fts0fts.cc103
-rw-r--r--storage/innobase/fts/fts0opt.cc54
-rw-r--r--storage/innobase/fts/fts0que.cc63
4 files changed, 167 insertions, 87 deletions
diff --git a/storage/innobase/fts/fts0ast.cc b/storage/innobase/fts/fts0ast.cc
index 14e52a99e18..8c542aab289 100644
--- a/storage/innobase/fts/fts0ast.cc
+++ b/storage/innobase/fts/fts0ast.cc
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 2007, 2015, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 2007, 2016, Oracle and/or its affiliates. All Rights Reserved.
This program is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free Software
@@ -161,7 +161,7 @@ fts_ast_create_node_term_for_parser(
/* '%' as first char is forbidden for LIKE in internal SQL parser;
'%' as last char is reserved for wildcard search;*/
- if (len == 0 || len > fts_max_token_size
+ if (len == 0 || len > FTS_MAX_WORD_LEN
|| ptr[0] == '%' || ptr[len - 1] == '%') {
return(NULL);
}
@@ -537,6 +537,36 @@ fts_ast_node_print(
fts_ast_node_print_recursive(node, 0);
}
+/** Check only union operation involved in the node
+@param[in] node ast node to check
+@return true if the node contains only union else false. */
+bool
+fts_ast_node_check_union(
+ fts_ast_node_t* node)
+{
+ if (node->type == FTS_AST_LIST
+ || node->type == FTS_AST_SUBEXP_LIST
+ || node->type == FTS_AST_PARSER_PHRASE_LIST) {
+
+ for (node = node->list.head; node; node = node->next) {
+ if (!fts_ast_node_check_union(node)) {
+ return(false);
+ }
+ }
+
+ } else if (node->type == FTS_AST_OPER
+ && (node->oper == FTS_IGNORE
+ || node->oper == FTS_EXIST)) {
+
+ return(false);
+ } else if (node->type == FTS_AST_TEXT) {
+ /* Distance or phrase search query. */
+ return(false);
+ }
+
+ return(true);
+}
+
/******************************************************************//**
Traverse the AST - in-order traversal, except for the FTX_EXIST and FTS_IGNORE
nodes, which will be ignored in the first pass of each level, and visited in a
diff --git a/storage/innobase/fts/fts0fts.cc b/storage/innobase/fts/fts0fts.cc
index 5edee71a13a..5faa68be0ea 100644
--- a/storage/innobase/fts/fts0fts.cc
+++ b/storage/innobase/fts/fts0fts.cc
@@ -208,19 +208,20 @@ struct fts_tokenize_param_t {
ulint add_pos; /*!< Added position for tokens */
};
-/****************************************************************//**
-Run SYNC on the table, i.e., write out data from the cache to the
+/** Run SYNC on the table, i.e., write out data from the cache to the
FTS auxiliary INDEX table and clear the cache at the end.
@param[in,out] sync sync state
@param[in] unlock_cache whether unlock cache lock when write node
@param[in] wait whether wait when a sync is in progress
+@param[in] has_dict whether has dict operation lock
@return DB_SUCCESS if all OK */
static
dberr_t
fts_sync(
fts_sync_t* sync,
bool unlock_cache,
- bool wait);
+ bool wait,
+ bool has_dict);
/****************************************************************//**
Release all resources help by the words rb tree e.g., the node ilist. */
@@ -1979,7 +1980,6 @@ fts_create_common_tables(
func_exit:
if (error != DB_SUCCESS) {
-
for (it = common_tables.begin(); it != common_tables.end();
++it) {
row_drop_table_for_mysql(
@@ -3648,7 +3648,7 @@ fts_add_doc_by_id(
DBUG_EXECUTE_IF(
"fts_instrument_sync_debug",
- fts_sync(cache->sync, true, true);
+ fts_sync(cache->sync, true, true, false);
);
DEBUG_SYNC_C("fts_instrument_sync_request");
@@ -3929,6 +3929,8 @@ fts_write_node(
doc_id_t first_doc_id;
char table_name[MAX_FULL_NAME_LEN];
+ ut_a(node->ilist != NULL);
+
if (*graph) {
info = (*graph)->info;
} else {
@@ -4446,7 +4448,7 @@ fts_sync_index(
ut_ad(rbt_validate(index_cache->words));
- error = fts_sync_write_words(sync->trx, index_cache, sync->unlock_cache);
+ error = fts_sync_write_words(trx, index_cache, sync->unlock_cache);
#ifdef FTS_DOC_STATS_DEBUG
/* FTS_RESOLVE: the word counter info in auxiliary table "DOC_ID"
@@ -4463,13 +4465,11 @@ fts_sync_index(
}
/** Check if index cache has been synced completely
-@param[in,out] sync sync state
@param[in,out] index_cache index cache
@return true if index is synced, otherwise false. */
static
bool
fts_sync_index_check(
- fts_sync_t* sync,
fts_index_cache_t* index_cache)
{
const ib_rbt_node_t* rbt_node;
@@ -4492,14 +4492,36 @@ fts_sync_index_check(
return(true);
}
-/*********************************************************************//**
-Commit the SYNC, change state of processed doc ids etc.
+/** Reset synced flag in index cache when rollback
+@param[in,out] index_cache index cache */
+static
+void
+fts_sync_index_reset(
+ fts_index_cache_t* index_cache)
+{
+ const ib_rbt_node_t* rbt_node;
+
+ for (rbt_node = rbt_first(index_cache->words);
+ rbt_node != NULL;
+ rbt_node = rbt_next(index_cache->words, rbt_node)) {
+
+ fts_tokenizer_word_t* word;
+ word = rbt_value(fts_tokenizer_word_t, rbt_node);
+
+ fts_node_t* fts_node;
+ fts_node = static_cast<fts_node_t*>(ib_vector_last(word->nodes));
+
+ fts_node->synced = false;
+ }
+}
+
+/** Commit the SYNC, change state of processed doc ids etc.
+@param[in,out] sync sync state
@return DB_SUCCESS if all OK */
static MY_ATTRIBUTE((nonnull, warn_unused_result))
dberr_t
fts_sync_commit(
-/*============*/
- fts_sync_t* sync) /*!< in: sync state */
+ fts_sync_t* sync)
{
dberr_t error;
trx_t* trx = sync->trx;
@@ -4550,6 +4572,8 @@ fts_sync_commit(
<< " ins/sec";
}
+ /* Avoid assertion in trx_free(). */
+ trx->dict_operation_lock_mode = 0;
trx_free_for_background(trx);
return(error);
@@ -4572,6 +4596,10 @@ fts_sync_rollback(
index_cache = static_cast<fts_index_cache_t*>(
ib_vector_get(cache->indexes, i));
+ /* Reset synced flag so nodes will not be skipped
+ in the next sync, see fts_sync_write_words(). */
+ fts_sync_index_reset(index_cache);
+
for (j = 0; fts_index_selector[j].value; ++j) {
if (index_cache->ins_graph[j] != NULL) {
@@ -4597,6 +4625,9 @@ fts_sync_rollback(
rw_lock_x_unlock(&cache->lock);
fts_sql_rollback(trx);
+
+ /* Avoid assertion in trx_free(). */
+ trx->dict_operation_lock_mode = 0;
trx_free_for_background(trx);
}
@@ -4605,13 +4636,15 @@ FTS auxiliary INDEX table and clear the cache at the end.
@param[in,out] sync sync state
@param[in] unlock_cache whether unlock cache lock when write node
@param[in] wait whether wait when a sync is in progress
+@param[in] has_dict whether has dict operation lock
@return DB_SUCCESS if all OK */
static
dberr_t
fts_sync(
fts_sync_t* sync,
bool unlock_cache,
- bool wait)
+ bool wait,
+ bool has_dict)
{
ulint i;
dberr_t error = DB_SUCCESS;
@@ -4640,6 +4673,12 @@ fts_sync(
DEBUG_SYNC_C("fts_sync_begin");
fts_sync_begin(sync);
+ /* When sync in background, we hold dict operation lock
+ to prevent DDL like DROP INDEX, etc. */
+ if (has_dict) {
+ sync->trx->dict_operation_lock_mode = RW_S_LATCH;
+ }
+
begin_sync:
if (cache->total_size > fts_max_cache_size) {
/* Avoid the case: sync never finish when
@@ -4676,7 +4715,7 @@ begin_sync:
ib_vector_get(cache->indexes, i));
if (index_cache->index->to_be_dropped
- || fts_sync_index_check(sync, index_cache)) {
+ || fts_sync_index_check(index_cache)) {
continue;
}
@@ -4691,6 +4730,7 @@ end_sync:
}
rw_lock_x_lock(&cache->lock);
+ sync->interrupted = false;
sync->in_progress = false;
os_event_set(sync->event);
rw_lock_x_unlock(&cache->lock);
@@ -4714,19 +4754,23 @@ FTS auxiliary INDEX table and clear the cache at the end.
@param[in,out] table fts table
@param[in] unlock_cache whether unlock cache when write node
@param[in] wait whether wait for existing sync to finish
+@param[in] has_dict whether has dict operation lock
@return DB_SUCCESS on success, error code on failure. */
dberr_t
fts_sync_table(
dict_table_t* table,
bool unlock_cache,
- bool wait)
+ bool wait,
+ bool has_dict)
{
dberr_t err = DB_SUCCESS;
ut_ad(table->fts);
- if (!dict_table_is_discarded(table) && table->fts->cache) {
- err = fts_sync(table->fts->cache->sync, unlock_cache, wait);
+ if (!dict_table_is_discarded(table) && table->fts->cache
+ && !dict_table_is_corrupted(table)) {
+ err = fts_sync(table->fts->cache->sync,
+ unlock_cache, wait, has_dict);
}
return(err);
@@ -5114,7 +5158,7 @@ fts_tokenize_document(
ut_a(doc->charset);
doc->tokens = rbt_create_arg_cmp(
- sizeof(fts_token_t), innobase_fts_text_cmp, (void*) doc->charset);
+ sizeof(fts_token_t), innobase_fts_text_cmp, (void*) doc->charset);
if (parser != NULL) {
fts_tokenize_param_t fts_param;
@@ -5809,6 +5853,8 @@ fts_update_doc_id(
if (error == DB_SUCCESS) {
dict_index_t* clust_index;
+ dict_col_t* col = dict_table_get_nth_col(
+ table, table->fts->doc_col);
ufield->exp = NULL;
@@ -5816,8 +5862,8 @@ fts_update_doc_id(
clust_index = dict_table_get_first_index(table);
- ufield->field_no = dict_col_get_clust_pos(
- &table->cols[table->fts->doc_col], clust_index);
+ ufield->field_no = dict_col_get_clust_pos(col, clust_index);
+ dict_col_copy_type(col, dfield_get_type(&ufield->new_val));
/* It is possible we update record that has
not yet be sync-ed from last crash. */
@@ -6784,11 +6830,7 @@ fts_fake_hex_to_dec(
#ifdef UNIV_DEBUG
ret =
#endif /* UNIV_DEBUG */
-#ifdef _WIN32
- sscanf(tmp_id, "%016llu", &dec_id);
-#else
- sscanf(tmp_id, "%016llu", &dec_id);
-#endif /* _WIN32 */
+ sscanf(tmp_id, "%016" UINT64scan, &dec_id);
ut_ad(ret == 1);
return dec_id;
@@ -7995,10 +8037,9 @@ func_exit:
consistent state. For now consistency is check only by ensuring
index->page_no != FIL_NULL
@param[out] base_table table has host fts index
-@param[in,out] trx trx handler
-@return true if check certifies auxillary tables are sane false otherwise. */
-bool
-fts_is_corrupt(
+@param[in,out] trx trx handler */
+void
+fts_check_corrupt(
dict_table_t* base_table,
trx_t* trx)
{
@@ -8016,7 +8057,7 @@ fts_is_corrupt(
fts_get_table_name(&fts_table, table_name);
dict_table_t* aux_table = dict_table_open_on_name(
- table_name, FALSE, FALSE, DICT_ERR_IGNORE_NONE);
+ table_name, true, FALSE, DICT_ERR_IGNORE_NONE);
if (aux_table == NULL) {
dict_set_corrupted(
@@ -8045,6 +8086,4 @@ fts_is_corrupt(
dict_table_close(aux_table, FALSE, FALSE);
}
-
- return(sane);
}
diff --git a/storage/innobase/fts/fts0opt.cc b/storage/innobase/fts/fts0opt.cc
index 89cdcf26591..5989aff83f4 100644
--- a/storage/innobase/fts/fts0opt.cc
+++ b/storage/innobase/fts/fts0opt.cc
@@ -602,7 +602,7 @@ fts_zip_read_word(
/* Finished decompressing block. */
if (zip->zp->avail_in == 0) {
- /* Free the block that's been decompressed. */
+ /* Free the block thats been decompressed. */
if (zip->pos > 0) {
ulint prev = zip->pos - 1;
@@ -1507,6 +1507,12 @@ fts_optimize_write_word(
fts_node_t* node = (fts_node_t*) ib_vector_get(nodes, i);
if (error == DB_SUCCESS) {
+ /* Skip empty node. */
+ if (node->ilist == NULL) {
+ ut_ad(node->ilist_size == 0);
+ continue;
+ }
+
error = fts_write_node(
trx, &graph, fts_table, word, node);
@@ -2635,7 +2641,6 @@ fts_optimize_remove_table(
/** Send sync fts cache for the table.
@param[in] table table to sync */
-UNIV_INTERN
void
fts_optimize_request_sync_table(
dict_table_t* table)
@@ -2650,7 +2655,7 @@ fts_optimize_request_sync_table(
/* FTS optimizer thread is already exited */
if (fts_opt_start_shutdown) {
- ib::info() << "Try to remove table " << table->name
+ ib::info() << "Try to sync table " << table->name
<< " after FTS optimize thread exiting.";
return;
}
@@ -2964,7 +2969,7 @@ fts_optimize_sync_table(
if (table) {
if (dict_table_has_fts_index(table) && table->fts->cache) {
- fts_sync_table(table, true, false);
+ fts_sync_table(table, true, false, true);
}
dict_table_close(table, FALSE, FALSE);
@@ -3122,26 +3127,7 @@ fts_optimize_thread(
ib_vector_get(tables, i));
if (slot->state != FTS_STATE_EMPTY) {
- dict_table_t* table = NULL;
-
- /*slot->table may be freed, so we try to open
- table by slot->table_id.*/
- table = dict_table_open_on_id(
- slot->table_id, FALSE,
- DICT_TABLE_OP_NORMAL);
-
- if (table) {
-
- if (dict_table_has_fts_index(table)) {
- fts_sync_table(table, false, true);
- }
-
- if (table->fts) {
- fts_free(table);
- }
-
- dict_table_close(table, FALSE, FALSE);
- }
+ fts_optimize_sync_table(slot->table_id);
}
}
}
@@ -3155,7 +3141,7 @@ fts_optimize_thread(
/* We count the number of threads in os_thread_exit(). A created
thread should always use that to exit and not use return() to exit. */
- os_thread_exit(NULL);
+ os_thread_exit();
OS_THREAD_DUMMY_RETURN;
}
@@ -3189,11 +3175,9 @@ fts_optimize_is_init(void)
return(fts_optimize_wq != NULL);
}
-/**********************************************************************//**
-Signal the optimize thread to prepare for shutdown. */
+/** Shutdown fts optimize thread. */
void
-fts_optimize_start_shutdown(void)
-/*=============================*/
+fts_optimize_shutdown()
{
ut_ad(!srv_read_only_mode);
@@ -3222,17 +3206,5 @@ fts_optimize_start_shutdown(void)
os_event_destroy(fts_opt_shutdown_event);
ib_wqueue_free(fts_optimize_wq);
-}
-
-/**********************************************************************//**
-Reset the work queue. */
-void
-fts_optimize_end(void)
-/*==================*/
-{
- ut_ad(!srv_read_only_mode);
-
- // FIXME: Potential race condition here: We should wait for
- // the optimize thread to confirm shutdown.
fts_optimize_wq = NULL;
}
diff --git a/storage/innobase/fts/fts0que.cc b/storage/innobase/fts/fts0que.cc
index 8abeb63f0a4..dee7c59a58b 100644
--- a/storage/innobase/fts/fts0que.cc
+++ b/storage/innobase/fts/fts0que.cc
@@ -153,6 +153,13 @@ struct fts_query_t {
bool multi_exist; /*!< multiple FTS_EXIST oper */
st_mysql_ftparser* parser; /*!< fts plugin parser */
+
+ /** limit value for the fts query */
+ ulonglong limit;
+
+ /** number of docs fetched by query. This is to restrict the
+ result with limit value */
+ ulonglong n_docs;
};
/** For phrase matching, first we collect the documents and the positions
@@ -2700,7 +2707,7 @@ fts_query_phrase_split(
/*****************************************************************//**
Text/Phrase search.
@return DB_SUCCESS or error code */
-static __attribute__((warn_unused_result))
+static MY_ATTRIBUTE((warn_unused_result))
dberr_t
fts_query_phrase_search(
/*====================*/
@@ -3209,6 +3216,11 @@ fts_query_filter_doc_ids(
ulint decoded = 0;
ib_rbt_t* doc_freqs = word_freq->doc_freqs;
+ if (query->limit != ULONG_UNDEFINED
+ && query->n_docs >= query->limit) {
+ return(DB_SUCCESS);
+ }
+
/* Decode the ilist and add the doc ids to the query doc_id set. */
while (decoded < len) {
ulint freq = 0;
@@ -3296,11 +3308,17 @@ fts_query_filter_doc_ids(
/* Add the word to the document's matched RB tree. */
fts_query_add_word_to_document(query, doc_id, word);
}
+
+ if (query->limit != ULONG_UNDEFINED
+ && query->limit <= ++query->n_docs) {
+ goto func_exit;
+ }
}
/* Some sanity checks. */
ut_a(doc_id == node->last_doc_id);
+func_exit:
if (query->total_size > fts_result_cache_limit) {
return(DB_FTS_EXCEED_RESULT_CACHE_LIMIT);
} else {
@@ -3904,19 +3922,24 @@ fts_query_can_optimize(
}
}
-/*******************************************************************//**
-FTS Query entry point.
+/** FTS Query entry point.
+@param[in] trx transaction
+@param[in] index fts index to search
+@param[in] flags FTS search mode
+@param[in] query_str FTS query
+@param[in] query_len FTS query string len in bytes
+@param[in,out] result result doc ids
+@param[in] limit limit value
@return DB_SUCCESS if successful otherwise error code */
dberr_t
fts_query(
-/*======*/
- trx_t* trx, /*!< in: transaction */
- dict_index_t* index, /*!< in: The FTS index to search */
- uint flags, /*!< in: FTS search mode */
- const byte* query_str, /*!< in: FTS query */
- ulint query_len, /*!< in: FTS query string len
- in bytes */
- fts_result_t** result) /*!< in/out: result doc ids */
+ trx_t* trx,
+ dict_index_t* index,
+ uint flags,
+ const byte* query_str,
+ ulint query_len,
+ fts_result_t** result,
+ ulonglong limit)
{
fts_query_t query;
dberr_t error = DB_SUCCESS;
@@ -3971,13 +3994,16 @@ fts_query(
if (flags & FTS_EXPAND) {
query.wildcard_words = rbt_create_arg_cmp(
- sizeof(fts_string_t), innobase_fts_text_cmp, (void*)charset);
+ sizeof(fts_string_t), innobase_fts_text_cmp, (void *)charset);
}
query.total_size += SIZEOF_RBT_CREATE;
query.total_docs = dict_table_get_n_rows(index->table);
+ query.limit = limit;
+
+ query.n_docs = 0;
#ifdef FTS_DOC_STATS_DEBUG
if (ft_enable_diag_print) {
error = fts_get_total_word_count(
@@ -4053,6 +4079,19 @@ fts_query(
fts_result_cache_limit = 2048;
);
+ /* Optimisation is allowed for limit value
+ when
+ i) No ranking involved
+ ii) Only FTS Union operations involved. */
+ if (query.limit != ULONG_UNDEFINED
+ && !fts_ast_node_check_union(ast)) {
+ query.limit = ULONG_UNDEFINED;
+ }
+
+ DBUG_EXECUTE_IF("fts_union_limit_off",
+ query.limit = ULONG_UNDEFINED;
+ );
+
/* Traverse the Abstract Syntax Tree (AST) and execute
the query. */
query.error = fts_ast_visit(
View Lorry Controller Status