Add support for PostgreSQL regex match

[b41sh]

Which issue does this PR close?

Closes #795

Rationale for this change

Implement PostgreSQL pattern matching operator

What changes are included in this PR?

1. modify sqlparser to support PostgreSQL regex match syntax sqlparser-rs#328
2. use arrow-rs kernel regexp_match function to do regular expression

Are there any user-facing changes?

support pattern matching using POSIX regular expressions

> CREATE EXTERNAL TABLE X STORED AS PARQUET LOCATION './benchmarks/parquet/nation/';

0 rows in set. Query took 0.005 seconds.
> select n_name from X where n_name ~ '^[A-D]';
+-----------+
| n_name    |
+-----------+
| ARGENTINA |
| BRAZIL    |
| CANADA    |
| CHINA     |
+-----------+
4 rows in set. Query took 0.010 seconds.
> select n_name from X where n_name ~* '^[a-d]';
+-----------+
| n_name    |
+-----------+
| ARGENTINA |
| BRAZIL    |
| CANADA    |
| CHINA     |
+-----------+
4 rows in set. Query took 0.008 seconds.
> select n_name from X where n_name !~ '^[C-R]';
+----------------+
| n_name         |
+----------------+
| ARGENTINA      |
| BRAZIL         |
| SAUDI ARABIA   |
| VIETNAM        |
| UNITED KINGDOM |
| UNITED STATES  |
+----------------+
6 rows in set. Query took 0.009 seconds.
> select n_name from X where n_name !~* '^[c-r]';
+----------------+
| n_name         |
+----------------+
| ARGENTINA      |
| BRAZIL         |
| SAUDI ARABIA   |
| VIETNAM        |
| UNITED KINGDOM |
| UNITED STATES  |
+----------------+
6 rows in set. Query took 0.009 seconds.

[alamb]

Thanks @b41sh -- I plan to review this PR tomorrow morning. I apologize for being backed up on review

[alamb]

I tested this PR out locally and it looks great 🥇 Nice work @b41sh! I found the code clean and beautiful to review.

The only other thing I can think of that might be worth adding is some end-to-end tests in datafusion/tests/sql.rs.

For anyone else interested, here is how I tested:

Get updated sqlparser

diff --git a/Cargo.toml b/Cargo.toml
index d6da8c14c..bb3aa2001 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -29,3 +29,6 @@ members = [
 ]

 exclude = ["python"]
+
+[patch.crates-io]
+sqlparser = { git = "https://github.com/b41sh/sqlparser-rs.git", branch = "regexp_match"}
\ No newline at end of file
diff --git a/ballista/rust/core/Cargo.toml b/ballista/rust/core/Cargo.toml
index b2fa50c88..66f61b6e6 100644
--- a/ballista/rust/core/Cargo.toml
+++ b/ballista/rust/core/Cargo.toml
@@ -37,7 +37,7 @@ hashbrown = "0.11"
 log = "0.4"
 prost = "0.8"
 serde = {version = "1", features = ["derive"]}
-sqlparser = "0.9.0"
+sqlparser = "0.9.1-alpha.0"
 tokio = "1.0"
 tonic = "0.5"
 uuid = { version = "0.8", features = ["v4"] }
diff --git a/datafusion/Cargo.toml b/datafusion/Cargo.toml
index 286be8a7a..9819275ab 100644
--- a/datafusion/Cargo.toml
+++ b/datafusion/Cargo.toml
@@ -51,7 +51,7 @@ ahash = "0.7"
 hashbrown = { version = "0.11", features = ["raw"] }
 arrow = { version = "5.1", features = ["prettyprint"] }
 parquet = { version = "5.1", features = ["arrow"] }
-sqlparser = "0.9.0"
+sqlparser = "0.9.1-alpha.0"
 paste = "^1.0"
 num_cpus = "1.13.0"
 chrono = "0.4"

Run datafusion-cli:

echo "foo" > /tmp/foo.csv
echo "Bar" >> /tmp/foo.csv
echo "Baz" >> /tmp/foo.csv
echo "ZZ" >> /tmp/foo.csv
cargo run -p datafusion-cli

> CREATE EXTERNAL TABLE foo(a VARCHAR) STORED AS CSV LOCATION '/tmp/foo.csv';
0 rows in set. Query took 0.002 seconds.


> select * from foo where a ~ 'z';
+-----+
| a   |
+-----+
| Baz |
+-----+
1 row in set. Query took 0.012 seconds.

> select * from foo where a ~* 'z';
+-----+
| a   |
+-----+
| Baz |
| ZZ  |
+-----+
2 rows in set. Query took 0.026 seconds.

> select * from foo where a !~ 'z';
+-----+
| a   |
+-----+
| foo |
| Bar |
| ZZ  |
+-----+
3 rows in set. Query took 0.012 seconds.

> select * from foo where a !~* 'z';
+-----+
| a   |
+-----+
| foo |
| Bar |
+-----+
2 rows in set. Query took 0.012 seconds.

[alamb]

It unfortunate that the arrow-rs doesn't have a regexp_filter style kernel that returns a BooleanArray --

https://sourcegraph.com/github.com/apache/arrow-rs/-/blob/arrow/src/compute/kernels/regexp.rs?L16

I filed apache/arrow-rs#697 and #905 to track that potential work, if we want to improve performance of these operators in future versions of DataFusion,

[alamb]

#934 is now merged, so upon rebase this PR should pass CI and be ready to go

[alamb]

The python test failures have been resolved. One more rebase and I bet this one can finally be merged in. Super kudos for keeping with it @b41sh -- thank you

[b41sh]

The python test failures have been resolved. One more rebase and I bet this one can finally be merged in. Super kudos for keeping with it @b41sh -- thank you

I have merged the master, PTAL, thanks for your review @alamb

[alamb]

I think it looks great @b41sh -- thank you for persevering!