Try this

.github/workflows/unique_python_files.yml

@@ -62,14 +62,14 @@ jobs:
       - name: Ingest
         run: |
           mkdir combined/
-          find dataset/ -name '*.parquet' | parallel -j 1 --xargs -n4 poetry run pypi-data run-sql ${{ github.workspace }}/sql/unique_python_files.prql --per-thread-output --output=parquet --threads=2 combined/{#}.parquet {}
+          find dataset/ -name '*.parquet' | parallel -j 1 --xargs -n4 poetry run pypi-data run-sql ${{ github.workspace }}/sql/unique_python_files.prql --output=parquet --threads=2 combined/{#}.parquet {}
           
       - name: List
         run: ls combined/
 
       - name: Combine
         run: |
-          poetry run pypi-data run-sql ${{ github.workspace }}/sql/unique_python_files_combine.prql --per-thread-output --output=parquet --memory=3 --threads=2 unique-python-files.parquet combined/*.parquet
+          poetry run pypi-data run-sql ${{ github.workspace }}/sql/unique_python_files_combine.prql --output=parquet --memory=3 --threads=2 unique-python-files.parquet combined/*.parquet
 
 #
 #      - name: Gets latest created release info

sql/unique_python_files.prql

@@ -1,13 +1,6 @@
 prql target:sql.duckdb
 
-let any_value = column -> s"any_value({column})"
-
 from (read_parquet $1)
-filter path ~= "\\.py$"
+filter s"suffix({path}, '.py')"
 filter skip_reason == ""
-select {hash, path}
-group {hash} (
-aggregate {
-  path = any_value(path)
-})
-sort {hash}
+select {s"distinct on({hash}) {hash}", path}

sql/unique_python_files_combine.prql

@@ -1,10 +1,5 @@
 prql target:sql.duckdb
 
-let any_value = column -> s"any_value({column})"
-
 from (read_parquet $1)
-select {hash, path}
-group {hash} (
-aggregate {
-  path = any_value(path)
-})
+select {s"distinct on({hash}) {hash}", path}
+