0ct0pu5/wesnoth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Completely rewrote the table sampling code...

Browse source 
...to work with multiple tables and cleaned it up significantly in the
process.
This commit is contained in:
Gregory Shikhman 2009-09-01 05:30:56 +00:00
parent e32d3dd0d0
commit b55784c099
1 changed files with 24 additions and 40 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

64
website/stats.wesnoth.org/dbsample.py
View file

@ -23,49 +23,33 @@ def sample(size):
conn = MySQLdb.connect(configuration.DB_HOSTNAME,configuration.DB_USERNAME,configuration.DB_PASSWORD,configuration.DB_NAME)
curs = conn.cursor()
#find the maximum game_id in the DB
curs.execute("SELECT MAX(game_id) FROM GAMES")
curs.execute("SELECT DISTINCT `table` FROM _wsviews")
results = curs.fetchall()
max_id = results[0][0]
#look for an existing table with this sample size and drop it if it exists, then create a new one
tblname = configuration.DB_TABLE_PREFIX+TBLSTRING+str(size)
curs.execute("SELECT count(*) FROM information_schema.tables WHERE table_schema = 'corn' AND table_name = '"+tblname+"'")
results = curs.fetchall()
exists = results[0][0] == 1
if exists:
curs.execute("DROP TABLE IF EXISTS "+tblname)
curs.execute("""
CREATE TABLE `"""+tblname+"""` (
`game_id` int(11) NOT NULL auto_increment,
`timestamp` datetime NOT NULL,
`user_id` char(32) NOT NULL,
`serial` char(18) NOT NULL,
`platform` char(8) default NULL,
`version` char(14) default NULL,
`campaign` char(40) default NULL,
`difficulty` char(20) default NULL,
`gold` int(11) default NULL,
`turns` int(11) default NULL,
`scenario` char(40) default NULL,
`start_turn` int(11) default NULL,
`time` int(11) default NULL,
`result` enum('victory','defeat','quit') default NULL,
`end_time` int(11) default NULL,
`end_gold` int(11) default NULL,
`end_turn` int(11) default NULL,
PRIMARY KEY (`game_id`)
) ENGINE=MyISAM AUTO_INCREMENT=2450740 DEFAULT CHARSET=utf8 """)
#randomly pick size number of entries from the main DB and put them into this sample
choices = random.sample(range(1,max_id),size)
for c in choices:
curs.execute("SELECT * FROM GAMES WHERE `game_id`=%s",c)
for tbl in results:
tbl = tbl[0]
sample_tbl = "%s%s%s%d" % (configuration.DB_TABLE_PREFIX,TBLSTRING,tbl,size)
curs.execute("DROP TABLE IF EXISTS %s" % (sample_tbl))
curs.execute("CREATE TABLE %s LIKE %s" % (sample_tbl,tbl))
curs.execute("SELECT MAX(game_id) FROM %s" % (tbl,)) #we assume game_id is always the primary key
results = curs.fetchall()
if len(results) != 0:
#print results[0]
curs.execute("""INSERT INTO %s (game_id,timestamp,user_id,serial,platform,version,
campaign,difficulty,gold,turns,scenario,start_turn,time,result,end_time,end_gold,end_turn) VALUES(%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s)""",results[0])
max_id = int(results[0][0])
if max_id < size:
size = max_id
choices = random.sample(range(0,max_id),size)
inserts = 0
for c in choices:
curs.execute("SELECT * FROM %s WHERE game_id=%d" % (tbl,c))
results = curs.fetchall()
if len(results) != 0:
query = "INSERT INTO %s SELECT * FROM %s WHERE game_id=%d" % (sample_tbl,tbl,c)
curs.execute(query)
conn.commit()
conn.close()
sample(10000)
sample(100000)