From 2e3328fce0643f0e4d50bcc4fe6969aaa8c3559a Mon Sep 17 00:00:00 2001
From: Harshal Sheth <hsheth2@gmail.com>
Date: Mon, 5 May 2025 09:24:50 -0700
Subject: [PATCH] chore(ingest): bump sqlglot dep (#13411)

---
 metadata-ingestion/setup.py                                    | 2 +-
 metadata-ingestion/src/datahub/sql_parsing/_sqlglot_patch.py   | 3 +--
 .../sql_parsing/goldens/test_postgres_update_subselect.json    | 2 +-
 .../goldens/test_snowflake_default_normalization.json          | 2 +-
 4 files changed, 4 insertions(+), 5 deletions(-)

diff --git a/metadata-ingestion/setup.py b/metadata-ingestion/setup.py
index ee28c2ead2..86e48c90f5 100644
--- a/metadata-ingestion/setup.py
+++ b/metadata-ingestion/setup.py
@@ -105,7 +105,7 @@ sqlglot_lib = {
     # We heavily monkeypatch sqlglot.
     # We used to maintain an acryl-sqlglot fork: https://github.com/tobymao/sqlglot/compare/main...hsheth2:sqlglot:main?expand=1
     # but not longer do.
-    "sqlglot[rs]==26.6.0",
+    "sqlglot[rs]==26.16.4",
     "patchy==2.8.0",
 }
 
diff --git a/metadata-ingestion/src/datahub/sql_parsing/_sqlglot_patch.py b/metadata-ingestion/src/datahub/sql_parsing/_sqlglot_patch.py
index af1a886549..88cc68ef85 100644
--- a/metadata-ingestion/src/datahub/sql_parsing/_sqlglot_patch.py
+++ b/metadata-ingestion/src/datahub/sql_parsing/_sqlglot_patch.py
@@ -163,8 +163,7 @@ def _patch_lineage() -> None:
 -    source_columns = set(find_all_in_scope(select, exp.Column))
 +    source_columns = list(find_all_in_scope(select, exp.Column))
 
--    # If the source is a UDTF find columns used in the UTDF to generate the table
-+    # If the source is a UDTF find columns used in the UDTF to generate the table
+     # If the source is a UDTF find columns used in the UDTF to generate the table
 +    source = scope.expression
      if isinstance(source, exp.UDTF):
 -        source_columns |= set(source.find_all(exp.Column))
diff --git a/metadata-ingestion/tests/unit/sql_parsing/goldens/test_postgres_update_subselect.json b/metadata-ingestion/tests/unit/sql_parsing/goldens/test_postgres_update_subselect.json
index 129912dd86..daf5a688de 100644
--- a/metadata-ingestion/tests/unit/sql_parsing/goldens/test_postgres_update_subselect.json
+++ b/metadata-ingestion/tests/unit/sql_parsing/goldens/test_postgres_update_subselect.json
@@ -19,7 +19,7 @@
                         "com.linkedin.pegasus2avro.schema.StringType": {}
                     }
                 },
-                "native_column_type": "VARCHAR"
+                "native_column_type": "VARCHAR(16777216)"
             },
             "upstreams": [
                 {
diff --git a/metadata-ingestion/tests/unit/sql_parsing/goldens/test_snowflake_default_normalization.json b/metadata-ingestion/tests/unit/sql_parsing/goldens/test_snowflake_default_normalization.json
index c56ce87261..0d742d712b 100644
--- a/metadata-ingestion/tests/unit/sql_parsing/goldens/test_snowflake_default_normalization.json
+++ b/metadata-ingestion/tests/unit/sql_parsing/goldens/test_snowflake_default_normalization.json
@@ -75,7 +75,7 @@
                         "com.linkedin.pegasus2avro.schema.NumberType": {}
                     }
                 },
-                "native_column_type": "DECIMAL"
+                "native_column_type": "DECIMAL(10, 2)"
             },
             "upstreams": [
                 {