mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-09-30 05:30:51 +00:00
b079dacfd1
* Added part_log * first_test * filter and hits_res * Add renamer and drawer * Add columns database and table into PartLog * Add normal way to get table_name and database_name from part * improve drawer * add stats for random size parts * Merge converter and drawer * make drawer more informative * add new data * add new data * new data * add long range stats * for checking best way * Add add_parts script * Good style for global merge * delete commented code * Fixed spaces to tabs * Note that Stopwatch is started automatically. * Style * Update StorageMergeTree.cpp * Update StorageReplicatedMergeTree.cpp * Switch act_time_ms to duration_ms * Added ability to disable part_log * fixed getPartLog * fix usage getPartLog * fix
62 lines
1.5 KiB
Python
62 lines
1.5 KiB
Python
import time
|
|
import ast
|
|
from datetime import datetime
|
|
|
|
FILE='data.tsv'
|
|
|
|
def get_metrix():
|
|
data = []
|
|
time_to_merge = 0
|
|
count_of_parts = 0
|
|
max_count_of_parts = 0
|
|
parts_in_time = []
|
|
last_date = 0
|
|
for line in open(FILE):
|
|
fields = line.split('\t')
|
|
last_date = datetime.strptime(fields[2], '%Y-%m-%d %H:%M:%S')
|
|
break
|
|
|
|
for line in open(FILE):
|
|
fields = line.split('\t')
|
|
cur_date = datetime.strptime(fields[2], '%Y-%m-%d %H:%M:%S')
|
|
if fields[0] == '2':
|
|
time_to_merge += int(fields[4])
|
|
list = ast.literal_eval(fields[-1])
|
|
count_of_parts -= len(list) - 1
|
|
else:
|
|
count_of_parts += 1
|
|
|
|
if max_count_of_parts < count_of_parts:
|
|
max_count_of_parts = count_of_parts
|
|
|
|
parts_in_time.append([(cur_date-last_date).total_seconds(), count_of_parts])
|
|
last_date = cur_date
|
|
|
|
stats_parts_in_time = []
|
|
global_time = 0
|
|
average_parts = 0
|
|
for i in range(max_count_of_parts + 1):
|
|
stats_parts_in_time.append(0)
|
|
|
|
for elem in parts_in_time:
|
|
stats_parts_in_time[elem[1]] += elem[0]
|
|
global_time += elem[0]
|
|
average_parts += elem[0] * elem[1]
|
|
|
|
for i in range(max_count_of_parts):
|
|
stats_parts_in_time[i] /= global_time
|
|
average_parts /= global_time
|
|
|
|
return time_to_merge, max_count_of_parts, average_parts, stats_parts_in_time
|
|
|
|
def main():
|
|
time_to_merge, max_parts, average_parts, stats_parts = get_metrix()
|
|
print('time_to_merge=', time_to_merge)
|
|
print('max_parts=', max_parts)
|
|
print('average_parts=', average_parts)
|
|
print('stats_parts=', stats_parts)
|
|
|
|
|
|
if __name__ == '__main__':
|
|
main()
|