diff --git a/swh/indexer/sql/50-func.sql b/swh/indexer/sql/50-func.sql --- a/swh/indexer/sql/50-func.sql +++ b/swh/indexer/sql/50-func.sql @@ -58,6 +58,7 @@ insert into content_mimetype (id, mimetype, encoding, indexer_configuration_id) select id, mimetype, encoding, indexer_configuration_id from tmp_content_mimetype tcm + order by id, indexer_configuration_id on conflict(id, indexer_configuration_id) do update set mimetype = excluded.mimetype, encoding = excluded.encoding; @@ -87,6 +88,7 @@ insert into content_language (id, lang, indexer_configuration_id) select id, lang, indexer_configuration_id from tmp_content_language tcl + order by id, indexer_configuration_id on conflict(id, indexer_configuration_id) do update set lang = excluded.lang; @@ -137,6 +139,7 @@ insert into content_ctags (id, name, kind, line, lang, indexer_configuration_id) select id, name, kind, line, lang, indexer_configuration_id from tmp_content_ctags tct + order by id, hash_sha1(name), kind, line, lang, indexer_configuration_id on conflict(id, hash_sha1(name), kind, line, lang, indexer_configuration_id) do nothing; @@ -218,6 +221,7 @@ (select id from fossology_license where name = tcl.license) as license, indexer_configuration_id from tmp_content_fossology_license tcl + order by tcl.id, license, indexer_configuration_id on conflict(id, license_id, indexer_configuration_id) do update set license_id = excluded.license_id; @@ -249,6 +253,7 @@ insert into content_metadata (id, metadata, indexer_configuration_id) select id, metadata, indexer_configuration_id from tmp_content_metadata tcm + order by id, indexer_configuration_id on conflict(id, indexer_configuration_id) do update set metadata = excluded.metadata; @@ -292,6 +297,7 @@ insert into directory_intrinsic_metadata (id, metadata, mappings, indexer_configuration_id) select id, metadata, mappings, indexer_configuration_id from tmp_directory_intrinsic_metadata tcm + order by id, indexer_configuration_id on conflict(id, indexer_configuration_id) do update set metadata = excluded.metadata, @@ -360,6 +366,7 @@ select id, metadata, indexer_configuration_id, from_directory, metadata_tsvector, mappings from tmp_origin_intrinsic_metadata + order by id, indexer_configuration_id on conflict(id, indexer_configuration_id) do update set metadata = excluded.metadata, @@ -433,6 +440,7 @@ select id, metadata, indexer_configuration_id, from_remd_id, metadata_tsvector, mappings from tmp_origin_extrinsic_metadata + order by id, indexer_configuration_id on conflict(id, indexer_configuration_id) do update set metadata = excluded.metadata, @@ -475,6 +483,7 @@ begin insert into indexer_configuration(tool_name, tool_version, tool_configuration) select tool_name, tool_version, tool_configuration from tmp_indexer_configuration tmp + order by tool_name, tool_version, tool_configuration on conflict(tool_name, tool_version, tool_configuration) do nothing; return query diff --git a/swh/indexer/storage/__init__.py b/swh/indexer/storage/__init__.py --- a/swh/indexer/storage/__init__.py +++ b/swh/indexer/storage/__init__.py @@ -294,7 +294,6 @@ cur=None, ) -> Dict[str, int]: check_id_duplicates(mimetypes) - mimetypes.sort(key=lambda m: m.id) self.journal_writer.write_additions("content_mimetype", mimetypes) db.mktemp_content_mimetype(cur) db.copy_to( @@ -342,7 +341,6 @@ cur=None, ) -> Dict[str, int]: check_id_duplicates(licenses) - licenses.sort(key=lambda m: m.id) self.journal_writer.write_additions("content_fossology_license", licenses) db.mktemp_content_fossology_license(cur) db.copy_to( @@ -407,7 +405,6 @@ cur=None, ) -> Dict[str, int]: check_id_duplicates(metadata) - metadata.sort(key=lambda m: m.id) self.journal_writer.write_additions("content_metadata", metadata) db.mktemp_content_metadata(cur) @@ -461,7 +458,6 @@ cur=None, ) -> Dict[str, int]: check_id_duplicates(metadata) - metadata.sort(key=lambda m: m.id) self.journal_writer.write_additions("directory_intrinsic_metadata", metadata) db.mktemp_directory_intrinsic_metadata(cur) @@ -505,7 +501,6 @@ cur=None, ) -> Dict[str, int]: check_id_duplicates(metadata) - metadata.sort(key=lambda m: m.id) self.journal_writer.write_additions("origin_intrinsic_metadata", metadata) db.mktemp_origin_intrinsic_metadata(cur) @@ -647,7 +642,6 @@ cur=None, ) -> Dict[str, int]: check_id_duplicates(metadata) - metadata.sort(key=lambda m: m.id) self.journal_writer.write_additions("origin_extrinsic_metadata", metadata) db.mktemp_origin_extrinsic_metadata(cur)