ARROW-12567: [C++][Gandiva] Implement ILIKE SQL function #10179

jvictorhuguenin · 2021-04-27T19:41:27Z

No description provided.

github-actions · 2021-04-27T19:41:59Z

https://issues.apache.org/jira/browse/ARROW-12567

projjal · 2021-04-28T08:08:43Z

cpp/src/gandiva/ilike_holder.h

+namespace gandiva {
+
+/// Function Holder for SQL 'ilike'
+class GANDIVA_EXPORT IlikeHolder : public FunctionHolder {


You can reuse the LikeHolder class and provide the case-sensitivity option in the make method

Ok, I created a Make method and constructor for the ILIKE function on LikeHolder, and deleted the IlikeHolder

projjal · 2021-04-28T08:11:25Z

cpp/src/gandiva/ilike_holder.cc

+    std::string& pattern = holder->pattern_;
+    auto literal_type = node.children().at(1)->return_type();
+
+    if (RE2::FullMatch(pattern, starts_with_regex_)) {


this seems incorrect, they should be case-insensitive

not really, the pattern has (?i)(google/re2#197) to turn case-insensitive on for the pattern.

this matching doesn't matter..actual starts_with, ends_with, sub_str are not case-insensitive

Here we match on the pattern for starts_with, ends_with cases, and replace with them..which won't be case-insensitive

Ok, now both like and ilike are using the same method.

This is not yet fixed. You are still replacing with starts_with, ends_with which are case-sensitive. You need to disable TryOptimise for ILIKE

Done, sorry I thought first that it had no difference.

projjal · 2021-04-30T06:14:00Z

cpp/src/gandiva/like_holder.h

@@ -48,6 +51,9 @@ class GANDIVA_EXPORT LikeHolder : public FunctionHolder {
 private:
  explicit LikeHolder(const std::string& pattern) : pattern_(pattern), regex_(pattern) {}

+  explicit LikeHolder(const std::string& pattern, RE2::Options regex_op)


You don't need explicit here. explicit is used to prevent implicit conversions.

projjal · 2021-04-30T06:21:20Z

cpp/src/gandiva/like_holder.cc

+    RE2::Options regex_op;
+    regex_op.set_case_sensitive(false);  // set case-insensitive for ilike function.
+
+    return Make(arrow::util::get<std::string>(literal->holder()), holder, regex_op);


Why create a new method for this. Modify the existing one take options

projjal · 2021-05-02T07:55:13Z

cpp/src/gandiva/like_holder.cc

  std::string pcre_pattern;
  ARROW_RETURN_NOT_OK(RegexUtil::SqlLikePatternToPcre(sql_pattern, pcre_pattern));

-  auto lholder = std::shared_ptr<LikeHolder>(new LikeHolder(pcre_pattern));
+  std::shared_ptr<LikeHolder> lholder;
+  if (regex_op.case_sensitive()) {


You can just call holder = std::shared_ptr<LikeHolder>(new LikeHolder(pcre_pattern, regex_op));
No need for if here

praveenbingo · 2021-06-07T10:28:56Z

@projjal needs a rebase

Closes apache#10179 from jvictorhuguenin/feature/implement-sql-ilike and squashes the following commits: f160880 <frank400> Optimize holder constructor call 97e6e2d <frank400> Remove unnecessary Make method c2363b1 <frank400> Disable TryOptimize for ilike a484149 <frank400> Fix checkstyle on cmake file c6a8372 <frank400> Delete unnecessary holder 4be6cc6 <frank400> Fix redefined function b78085a <frank400> Fix miss include 2efd43e <frank400> Implement ilike function Authored-by: frank400 <j.victorhuguenin2018@gmail.com> Signed-off-by: Praveen <praveen@dremio.com> (cherry picked from commit 0072c67)

Closes apache#10179 from jvictorhuguenin/feature/implement-sql-ilike and squashes the following commits: f160880 <frank400> Optimize holder constructor call 97e6e2d <frank400> Remove unnecessary Make method c2363b1 <frank400> Disable TryOptimize for ilike a484149 <frank400> Fix checkstyle on cmake file c6a8372 <frank400> Delete unnecessary holder 4be6cc6 <frank400> Fix redefined function b78085a <frank400> Fix miss include 2efd43e <frank400> Implement ilike function Authored-by: frank400 <j.victorhuguenin2018@gmail.com> Signed-off-by: Praveen <praveen@dremio.com>

* ARROW-11960: [C++][Gandiva] Support escape in LIKE Add gdv_fn_like_utf8_utf8_int8 function in Gandiva to support escape char in LIKE. An escape char is stored in an int8 type which is compatible with char type in C++. Closes apache#9700 from Crystrix/arrow-11960 Authored-by: crystrix <chenxi.li@live.com> Signed-off-by: Sutou Kouhei <kou@clear-code.com> * ARROW-12567: [C++][Gandiva] Implement ILIKE SQL function Closes apache#10179 from jvictorhuguenin/feature/implement-sql-ilike and squashes the following commits: f160880 <frank400> Optimize holder constructor call 97e6e2d <frank400> Remove unnecessary Make method c2363b1 <frank400> Disable TryOptimize for ilike a484149 <frank400> Fix checkstyle on cmake file c6a8372 <frank400> Delete unnecessary holder 4be6cc6 <frank400> Fix redefined function b78085a <frank400> Fix miss include 2efd43e <frank400> Implement ilike function Authored-by: frank400 <j.victorhuguenin2018@gmail.com> Signed-off-by: Praveen <praveen@dremio.com> * ARROW-12410: [C++][Gandiva] Implement regexp_replace function on Gandiva Closes apache#10059 from rodrigojdebem/feature/implement-regexp-replace and squashes the following commits: baf2778 <rodrigojdebem> Add implementation for REGEXP_REPLACE Authored-by: rodrigojdebem <rodrigodebem1@gmail.com> Signed-off-by: Praveen <praveen@dremio.com> Co-authored-by: crystrix <chenxi.li@live.com> Co-authored-by: frank400 <j.victorhuguenin2018@gmail.com> Co-authored-by: rodrigojdebem <rodrigodebem1@gmail.com>

github-actions bot added Component: Gandiva Component: C++ labels Apr 27, 2021

projjal reviewed Apr 28, 2021

View reviewed changes

projjal reviewed Apr 30, 2021

View reviewed changes

projjal reviewed May 2, 2021

View reviewed changes

projjal approved these changes May 12, 2021

View reviewed changes

jvictorhuguenin added 8 commits June 7, 2021 09:50

Implement ilike function

2efd43e

Fix miss include

b78085a

Fix redefined function

4be6cc6

Delete unnecessary holder

c6a8372

Fix checkstyle on cmake file

a484149

Disable TryOptimize for ilike

c2363b1

Remove unnecessary Make method

97e6e2d

Optimize holder constructor call

f160880

jvictorhuguenin force-pushed the feature/implement-sql-ilike branch from 608e08f to f160880 Compare June 7, 2021 13:00

praveenbingo closed this in 0072c67 Jul 6, 2021

anthonylouisbsb deleted the feature/implement-sql-ilike branch September 28, 2021 22:32

asfimport mentioned this pull request Jul 6, 2021

[C++][Gandiva] Implement LPAD and RPAD functions for string input values #28326

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

ARROW-12567: [C++][Gandiva] Implement ILIKE SQL function #10179

ARROW-12567: [C++][Gandiva] Implement ILIKE SQL function #10179

jvictorhuguenin commented Apr 27, 2021

github-actions bot commented Apr 27, 2021

projjal Apr 28, 2021

jvictorhuguenin Apr 28, 2021 •

edited

Loading

projjal Apr 28, 2021

jvictorhuguenin Apr 28, 2021 •

edited

Loading

projjal Apr 28, 2021

projjal Apr 28, 2021

jvictorhuguenin Apr 28, 2021

projjal Apr 29, 2021

jvictorhuguenin Apr 29, 2021

projjal Apr 30, 2021

jvictorhuguenin Apr 30, 2021

projjal Apr 30, 2021

jvictorhuguenin Apr 30, 2021

projjal May 2, 2021

jvictorhuguenin May 3, 2021

praveenbingo commented Jun 7, 2021

ARROW-12567: [C++][Gandiva] Implement ILIKE SQL function #10179

ARROW-12567: [C++][Gandiva] Implement ILIKE SQL function #10179

Conversation

jvictorhuguenin commented Apr 27, 2021

github-actions bot commented Apr 27, 2021

Choose a reason for hiding this comment

jvictorhuguenin Apr 28, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jvictorhuguenin Apr 28, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

praveenbingo commented Jun 7, 2021

jvictorhuguenin Apr 28, 2021 •

edited

Loading

jvictorhuguenin Apr 28, 2021 •

edited

Loading