Fix batching with negatives cartodb_id

This commit is contained in:
Juan Ignacio Sánchez Lara
2018-07-11 12:28:39 +02:00
parent d46d51c3bb
commit da78b0bc65
3 changed files with 21 additions and 14 deletions

View File

@@ -6,6 +6,8 @@ DECLARE
enough_quota boolean;
remaining_quota integer;
max_batch_size integer;
max_cartodb_id integer;
min_cartodb_id integer;
cartodb_id_batch integer;
batches_n integer;
@@ -31,7 +33,8 @@ BEGIN
batch_size := MAX_SAFE_BATCH_SIZE;
END IF;
EXECUTE format('SELECT COUNT(1) from (%s) _x', query) INTO query_row_count;
EXECUTE format('SELECT count(1), max(cartodb_id), min(cartodb_id), ceil((max(cartodb_id) + 1 - min(cartodb_id))::float/%s) FROM (%s) _x', batch_size, query)
INTO query_row_count, max_cartodb_id, min_cartodb_id, batches_n;
RAISE DEBUG 'cdb_bulk_geocode_street_point --> query_row_count: %; query: %; country: %; state: %; city: %; street: %',
query_row_count, query, country_column, state_column, city_column, street_column;
@@ -40,8 +43,6 @@ BEGIN
RAISE EXCEPTION 'Remaining quota: %. Estimated cost: %', remaining_quota, query_row_count;
END IF;
EXECUTE format('SELECT ceil(max(cartodb_id)::float/%s) FROM (%s) _x', batch_size, query) INTO batches_n;
RAISE DEBUG 'batches_n: %', batches_n;
temp_table_name := 'bulk_geocode_street_' || md5(random()::text);
@@ -63,13 +64,13 @@ BEGIN
'WITH geocoding_data as (' ||
' SELECT ' ||
' json_build_object(''id'', cartodb_id, ''address'', %s, ''city'', %s, ''state'', %s, ''country'', %s) as data , ' ||
' floor((cartodb_id-1)::float/$1) as batch' ||
' floor((cartodb_id - $1)::float/$2) as batch' ||
' FROM (%s) _x' ||
') ' ||
'INSERT INTO %s SELECT (cdb_dataservices_client._cdb_bulk_geocode_street_point(jsonb_agg(data))).* ' ||
'FROM geocoding_data ' ||
'WHERE batch = $2', street_column, city_column, state_column, country_column, query, temp_table_name)
USING batch_size, cartodb_id_batch;
'WHERE batch = $3', street_column, city_column, state_column, country_column, query, temp_table_name)
USING min_cartodb_id, batch_size, cartodb_id_batch;
GET DIAGNOSTICS current_row_count = ROW_COUNT;
RAISE DEBUG 'Batch % --> %', cartodb_id_batch, current_row_count;