New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

LIKE implementation #476 #512

Merged

bluestreak01 merged 6 commits into questdb:master from mick2004:master

Aug 26, 2020

Contributor

mick2004 commented Jul 19, 2020

Tested via webUI and automated tests.

mick2004 added 3 commits

July 14, 2020 01:40


          Error while running JMH benchmark questdb#503

5ea28d1


          Text Function !~ doesnt work when used with single character questdb#505

f4e5396


          LIKE implementation questdb#476

fc0ad44

mick2004 requested review from bluestreak01, jaugsburger, mpsq, patrickSpaceSurfer and TheTanc as code owners

July 19, 2020 14:46

Member

bluestreak01 commented Jul 19, 2020

Thank you Abhishek!

Please have a look here on how “like” is expected to work:

https://www.w3schools.com/sql/trysql.asp?filename=trysql_select_like

Contributor Author

mick2004 commented Jul 19, 2020

@bluestreak01 sure

Member

bluestreak01 commented Jul 19, 2020

cool, thanks. I just wanted to draw your attention to the fact that 'like' is a pattern matching operator and like 'a' should only match exact a strings

Contributor Author

mick2004 commented Jul 20, 2020 •

edited

cool, thanks. I just wanted to draw your attention to the fact that 'like' is a pattern matching operator and like 'a' should only match exact a strings

@bluestreak01 Understood.So are below criteria ok for LIKE or I am missing anything ?

1.like '<>' return exact match of <>
2.like '<>%' return records starting with <>
3.like '%<>' return records ending with <>
4.like '%<>%' return match of <> anywhere.This is equivalent of ~= behaviour

Member

bluestreak01 commented Jul 20, 2020 •

edited

There is also underscore (_) - matches any one character. “Like” is best described here: https://www.w3schools.com/sql/sql_like.asp

% is a wildcard, it can be present in the pattern 0 or more types, more than 2 certainly. One way to deal with this would be rewriting “like” pattern to a regex and executing the latter

Contributor Author

mick2004 commented Jul 20, 2020

There is also underscore (_) - matches any one character. “Like” is best described here: https://www.w3schools.com/sql/sql_like.asp

% is a wildcard, it can be present in the pattern 0 or more types, more than 2 certainly. One way to deal with this would be rewriting “like” pattern to a regex and executing the latter

Sure,let me check in more depth.


          LIKE implementation questdb#476

6b5e6e2

Contributor Author

mick2004 commented Jul 21, 2020 •

edited

@bluestreak01 Pushed changes for:
1.Handling %
2.Handling _
3.Handling Exact match
4.Skipping special characters

Please review.

bluestreak01 requested changes

View reviewed changes

core/src/main/java/io/questdb/griffin/engine/functions/regex/LikeCharFunctionFactory.java Outdated

+                      @Override
+                      public boolean getBool(Record rec) {
+                          CharSequence cs = getArg().getStr(rec);
+                          return cs != null && cs.length()==1 && Character.toLowerCase(cs.charAt(0))==Character.toLowerCase(expected);

Member

bluestreak01 Jul 22, 2020

SQL 'like' is case sensitive

Contributor Author

mick2004 Jul 29, 2020

@bluestreak01 https://www.w3schools.com/sql/trysql.asp?filename=trysql_select_like

Try query : SELECT * FROM Customers WHERE CustomerName LIKE 'alfreds%';

I am getting results for Alfred. Just want to double check before making change.

core/src/main/java/io/questdb/griffin/engine/functions/regex/LikeStrFunctionFactory.java Outdated

+                      CharSequence likeString=args.getQuick(1).getStr(null);
+                      if (likeString == null) {
+                          throw SqlException.$(args.getQuick(1).getPosition(), "NULL likeString");

Member

bluestreak01 Jul 22, 2020

This message is shown to user as well as position of the error. User needs to know what to do with this. 'likeString' is not a commonly known term :)

Contributor Author

mick2004 Jul 29, 2020 •

edited

@bluestreak01 What should be message. Is "NULL not supported for like expression" ok ? else suggest some message

core/src/main/java/io/questdb/griffin/engine/functions/regex/LikeStrFunctionFactory.java Outdated

+                          return value;
+                      }
+                  }
+                  }

Member

bluestreak01 Jul 22, 2020

What IDE are you using? Would it be possible to format the code?

Contributor Author

mick2004 Jul 29, 2020

@bluestreak01 sure

core/src/main/java/io/questdb/griffin/engine/functions/regex/LikeStrFunctionFactory.java Outdated

+                          return "";
+                      }
+                      StringBuilder sb = new StringBuilder(s.length() * 2);

Member

bluestreak01 Jul 22, 2020

StringBuilder is transient and throw-away instance. Please instead use

Misc.getThreadLocalBuilder()

core/src/main/java/io/questdb/griffin/engine/functions/regex/LikeStrFunctionFactory.java Outdated

+                      }
+                      String regex = escapeSpecialChars(likeString).
+                              replace('_', '.').replace("%", ".*?");

Member

bluestreak01 Jul 22, 2020

this can be done in the same loop as escaping special characters, which would eliminate two extra loops and two string instances.

core/src/main/java/io/questdb/griffin/engine/functions/regex/LikeStrFunctionFactory.java Outdated

+                      try {
+                          Matcher matcher = Pattern.compile(regex,
+                                  Pattern.CASE_INSENSITIVE | Pattern.DOTALL).matcher("");

Member

bluestreak01 Jul 22, 2020

'like' is case sensitive

core/src/main/java/io/questdb/griffin/engine/functions/regex/LikeStrFunctionFactory.java Outdated

+                          }
+                          sb.append(c);
+                      }
+                      return sb.toString();

Member

bluestreak01 Jul 22, 2020

please use Chars.toString(). Instances of strings are liable to be cached in the future.

core/src/main/java/io/questdb/griffin/engine/functions/regex/LikeStrFunctionFactory.java Outdated

+                      int len = s.length();
+                      if (len == 0)
+                      {
+                          return "";

Member

bluestreak01 Jul 22, 2020

there is no unit test that hits this line

core/src/main/java/io/questdb/griffin/engine/functions/regex/LikeStrFunctionFactory.java Outdated



		} catch (PatternSyntaxException e) {
		throw SqlException.$(args.getQuick(1).getPosition() + e.getIndex() + 1, e.getMessage());

Member

bluestreak01 Jul 22, 2020

this code is untested

bluestreak01 reviewed

View reviewed changes

core/src/main/java/io/questdb/griffin/OperatorExpression.java Outdated

@@ @@ -54,6 +54,7 @@ @@
                       add(new OperatorExpression("and", 11, true, BINARY, false));
                       add(new OperatorExpression("or", 11, true, BINARY, false));
                       add(new OperatorExpression("not", 11, true, UNARY, false));
+                      add(new OperatorExpression("like", 7, true, BINARY));

Member

bluestreak01 Jul 22, 2020

'like' must not be a symbol (see 'and', 'not' etc.)

The reason is that symbols break up words, for example 'dislike' will be tokenized as 'dis' + 'like' if 'like' was a symbol. When 'like' is not a symbol - 'dislike' will remain whole and 'like' will only be tokenized as such when it is separated with whitespace, e.g. x like '%a'

Contributor Author

mick2004 Jul 29, 2020

@bluestreak01 understood

CLAassistant commented Aug 19, 2020 •

edited

All committers have signed the CLA.

mick2004 added 2 commits

August 20, 2020 00:00


          LIKE implementation questdb#476

ff96cc9


          latest sync

e776bc9

Contributor Author

mick2004 commented Aug 19, 2020

@bluestreak01 Can you please review again.Pushed changes as per your feedback.

mick2004 requested a review from bluestreak01

August 19, 2020 17:00

bluestreak01 approved these changes

View reviewed changes

bluestreak01 merged commit 0cae67f into questdb:master

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

bluestreak01 bluestreak01 approved these changes

jaugsburger Awaiting requested review from jaugsburger

mpsq Awaiting requested review from mpsq

patrickSpaceSurfer Awaiting requested review from patrickSpaceSurfer

TheTanc Awaiting requested review from TheTanc