From 3e9f39e8bb3d41c3354185de2dc624ff806a4fcd Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Philipp=20Gro=C3=9Fer?= Date: Fri, 10 Apr 2026 10:49:42 +0200 Subject: [PATCH] fixed shared hashing that caused memory issues --- transformplan/backends/duckdb.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/transformplan/backends/duckdb.py b/transformplan/backends/duckdb.py index 69a5a90..9c289d0 100644 --- a/transformplan/backends/duckdb.py +++ b/transformplan/backends/duckdb.py @@ -96,8 +96,8 @@ def compute_hash(self, data: duckdb.DuckDBPyRelation) -> str: f"COALESCE({_q(c)}::VARCHAR, '')" for c in cols ) sql = ( - f"SELECT md5(string_agg(row_str, '\\n' ORDER BY row_str)) AS h " - f"FROM (SELECT {concat_expr} AS row_str " + f"SELECT md5(string_agg(row_hash, '' ORDER BY row_hash)) AS h " + f"FROM (SELECT md5({concat_expr}) AS row_hash " f"FROM (SELECT {col_list} FROM {_sub(data)}) AS _s) AS _r" ) result = self._con.sql(sql).fetchone()