From ea977948d711b0d8446fe4d9ca90531df79123c4 Mon Sep 17 00:00:00 2001
From: Phillip Cloud <417981+cpcloud@users.noreply.github.com>
Date: Mon, 26 Aug 2024 09:57:19 -0400
Subject: [PATCH] fix(bigquery): disallow column names longer than 300
 characters (#9916)

Closes #8931.
---
 .../bigquery/tests/unit/test_compiler.py      |  9 +++++++++
 .../sql/compilers/bigquery/__init__.py        | 20 ++++++++++++++++++-
 2 files changed, 28 insertions(+), 1 deletion(-)

diff --git a/ibis/backends/bigquery/tests/unit/test_compiler.py b/ibis/backends/bigquery/tests/unit/test_compiler.py
index eb28db2b840f..e058cd214c86 100644
--- a/ibis/backends/bigquery/tests/unit/test_compiler.py
+++ b/ibis/backends/bigquery/tests/unit/test_compiler.py
@@ -693,3 +693,12 @@ def test_approx_quantiles(alltypes, quantiles, snapshot):
     query = alltypes.double_col.approx_quantile(quantiles).name("qs")
     result = ibis.to_sql(query, dialect="bigquery")
     snapshot.assert_match(result, "out.sql")
+
+
+def test_unreasonably_long_name():
+    expr = ibis.literal("hello, world!").name("a" * 301)
+    with pytest.raises(
+        com.IbisError,
+        match="BigQuery does not allow column names longer than 300 characters",
+    ):
+        ibis.to_sql(expr, dialect="bigquery")
diff --git a/ibis/backends/sql/compilers/bigquery/__init__.py b/ibis/backends/sql/compilers/bigquery/__init__.py
index 97ccdc1d6511..e803734d1f59 100644
--- a/ibis/backends/sql/compilers/bigquery/__init__.py
+++ b/ibis/backends/sql/compilers/bigquery/__init__.py
@@ -924,7 +924,25 @@ def visit_HashBytes(self, op, *, arg, how):
 
     @staticmethod
     def _gen_valid_name(name: str) -> str:
-        return "_".join(map(str.strip, _NAME_REGEX.findall(name))) or "tmp"
+        candidate = "_".join(map(str.strip, _NAME_REGEX.findall(name))) or "tmp"
+        # column names cannot be longer than 300 characters
+        #
+        # https://cloud.google.com/bigquery/docs/schemas#column_names
+        #
+        # it's easy to rename columns, so raise an exception telling the user
+        # to do so
+        #
+        # we could potentially relax this and support arbitrary-length columns
+        # by compressing the information using hashing, but there's no reason
+        # to solve that problem until someone encounters this error and cannot
+        # rename their columns
+        limit = 300
+        if len(candidate) > limit:
+            raise com.IbisError(
+                f"BigQuery does not allow column names longer than {limit:d} characters. "
+                "Please rename your columns to have fewer characters."
+            )
+        return candidate
 
     def visit_CountStar(self, op, *, arg, where):
         if where is not None: