def non_indexed_doc_ids_not_planned_to_index(
document_type: DocumentType,
pack_size: int = 100) -> Generator[List[int], None, None]:
table_name = doc_fields_table_name(document_type.code)
with connection.cursor() as cursor:
# return documents of the specified type which
# - do not exist in the corresponding fields cache
# - have no planned but not-started reindex tasks on them
cursor.execute(
'select dd.id \n'
'from document_document dd \n'
'left outer join "{table_name}" df on dd.id = df.document_id \n'
'left outer join lateral (select jsonb_array_elements(args->0) doc_id \n'
' from task_task \n'
' where name = %s \n'
' and own_status = %s\n'
' and date_work_start is null) tt on tt.doc_id = to_jsonb(dd.id) \n'
'where dd.document_type_id = %s and df.document_id is null and tt.doc_id is null'
.format(table_name=table_name),
[_get_reindex_task_name(), 'PENDING', document_type.uid])
rows = cursor.fetchmany(pack_size)
while rows:
yield [row[0] for row in rows]
rows = cursor.fetchmany(pack_size)
def do_migrate(apps, schema_editor):
DocumentType = apps.get_model('document', 'DocumentType')
with connection.cursor() as cursor:
for document_type in DocumentType.objects.all():
table_name = doc_fields_table_name(
document_type_code=document_type.code)
cursor.execute('DROP TABLE IF EXISTS "{table_name}"'.format(
table_name=table_name))
def there_are_non_indexed_docs_not_planned_to_index(
document_type: DocumentType, log: ProcessLogger) -> bool:
for doc_id in non_indexed_doc_ids_not_planned_to_index(document_type, 1):
if doc_id:
task_name = _get_reindex_task_name()
fields_table = doc_fields_table_name(document_type.code)
log.info(
f'there_are_non_indexed_docs_not_planned_to_index: '
f'found document id={doc_id} of type {document_type.code}, '
f'task {task_name}. Fields table: {fields_table}')
return True
return False