diff --git a/pkg/ccl/logictestccl/testdata/logic_test/partitioning b/pkg/ccl/logictestccl/testdata/logic_test/partitioning
index d4b1bb4b4ea3..1ca7eaba4573 100644
--- a/pkg/ccl/logictestccl/testdata/logic_test/partitioning
+++ b/pkg/ccl/logictestccl/testdata/logic_test/partitioning
@@ -46,6 +46,13 @@ CREATE TABLE t (a INT, b INT, c INT, PRIMARY KEY (a, b, c)) PARTITION BY LIST (a
     )
 )
 
+# Regression test for #37682
+statement error declared partition columns \(a, b\) do not match first 1 columns in index being partitioned \(rowid\)
+CREATE TABLE t (a INT DEFAULT 12, b INT DEFAULT 34) PARTITION BY LIST (a, b) (
+  PARTITION p1 VALUES IN ((1,2)),
+  PARTITION default VALUES IN (DEFAULT)
+)
+
 statement error PARTITION p1: name must be unique \(used twice in index "primary"\)
 CREATE TABLE t (a INT, b INT, c INT, PRIMARY KEY (a, b)) PARTITION BY LIST (a) (
     PARTITION p1 VALUES IN (1),
diff --git a/pkg/ccl/partitionccl/partition.go b/pkg/ccl/partitionccl/partition.go
index 2a793a0ca0a1..e07e85e0d815 100644
--- a/pkg/ccl/partitionccl/partition.go
+++ b/pkg/ccl/partitionccl/partition.go
@@ -178,7 +178,9 @@ func createPartitioningImpl(
 		}
 		cols = append(cols, col)
 		if string(partBy.Fields[i]) != col.Name {
-			n := colOffset + len(partBy.Fields)
+			// This used to print the first `colOffset + len(partBy.Fields)` fields
+			// but there might not be this many columns in the index. See #37682.
+			n := colOffset + i + 1
 			return partDesc, fmt.Errorf(
 				"declared partition columns (%s) do not match first %d columns in index being partitioned (%s)",
 				partitioningString(), n, strings.Join(indexDesc.ColumnNames[:n], ", "))