-
Notifications
You must be signed in to change notification settings - Fork 0
/
preload.py
139 lines (109 loc) · 4.91 KB
/
preload.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
from pyspark.sql import SparkSession, SQLContext
from pyspark import SparkConf
# from pyspark.sql import SQLContext
from structure import *
from query import *
from datetime import datetime
import properties as p
import argparse
spark, sc = None, None
schema_entities = {}
def init_table(baseurl):
date_format = '%Y-%m-%d'
customer, lineitem, nation, region, region, part, partsupp, orders, supplier = None, None, None, None, None, None, None, None, None
customer = sc.textFile("%s/customer.tbl" % baseurl) \
.map(lambda row: row.split('|')) \
.map(lambda row: [int(row[0])] + row[1:3] + [int(row[3])] + [row[4]] + [float(row[5])] + row[6:-1])
lineitem = sc.textFile("%s/lineitem.tbl" % baseurl) \
.map(lambda row: row.split('|')) \
.map(lambda row: [int(x) for x in row[0:4]] + [float(x) for x in row[4:8]] + row[8:10] + [datetime.strptime(x, date_format) for x in row[10:13]] + row[13:-1])
nation = sc.textFile("%s/nation.tbl" % baseurl) \
.map(lambda row: row.split('|')) \
.map(lambda row: [int(row[0]), row[1], int(row[2]), row[3]])
region = sc.textFile("%s/region.tbl" % baseurl) \
.map(lambda row: row.split('|')) \
.map(lambda row: [int(row[0]), row[1], row[2]])
part = sc.textFile("%s/part.tbl" % baseurl) \
.map(lambda row: row.split('|')) \
.map(lambda row: [int(row[0])] + row[1:5] + [int(row[5]), row[6], float(row[7]), row[8]])
partsupp = sc.textFile("%s/partsupp.tbl" % baseurl) \
.map(lambda row: row.split('|')) \
.map(lambda row: [int(x) for x in row[0:3]] + [float(row[3])] + [row[4]])
orders = sc.textFile("%s/orders.tbl" % baseurl) \
.map(lambda row: row.split('|')) \
.map(lambda row: [int(x) for x in row[0:2]] + [row[2]] + [float(row[3])] + [datetime.strptime(row[4], date_format)] + row[5:7] + [int(row[7])] + [row[8]])
supplier = sc.textFile("%s/supplier.tbl" % baseurl) \
.map(lambda row: row.split('|')) \
.map(lambda row: [int(row[0])] + row[1:3] + [int(row[3])] + [row[4]] + [float(row[5])] + [row[6]])
return customer, lineitem, nation, region, region, part, partsupp, orders, supplier
def init_spark(sp=None):
global spark, sc, schema_entities
if sp:
spark = sp
else:
conf = (SparkConf().setAppName("assignment"))
# conf.set("spark.driver.memory", "256g")
# conf.set("spark.executor.memory", "128g")
conf.set("spark.ui.port", p.port)
# conf.set("spark.sql.shuffle.partitions", "100")
spark = SparkSession.builder.config(conf=conf).getOrCreate()
sc = spark.sparkContext
customer, lineitem, nation, region, region, part, partsupp, orders, supplier = init_table(p.baseurl)
schema_entities = {
'customer': (customer, s_customer),
'lineitem': (lineitem, s_lineitem),
'nation': (nation, s_nation),
'region': (region, s_region),
'part': (part, s_part),
'partsupp': (partsupp, s_partsupp),
'orders': (orders, s_orders),
'supplier': (supplier, s_supplier)
}
def init_temple_dataframe():
for key, value in schema_entities.iteritems():
if value[0]:
df = spark.createDataFrame(value[0], value[1])
df.registerTempTable(key)
# check dataframe
# print(df.first())
def create_parquet(spark, prefix=''):
for key, value in schema_entities.iteritems():
if value[0]:
df = spark.createDataFrame(value[0], value[1])
# check dataframe
# print(df.first())
# save dataframe to parquet to save time
df.write.format("parquet").save("%s%s.parquet" % (prefix, key), mode='overwrite')
def load_parquet(spark, prefix='', tables=[]):
for x in tables:
df = spark.read.load("%s%s.parquet" % (prefix, x))
df.registerTempTable(x)
def remove_parquet(spark, tables):
for x in tables:
spark.catalog.dropTempView(x)
def cache(spark, tables, prefix=''):
for x in tables:
df = spark.read.load("%s%s.parquet" % (prefix, x))
df.registerTempTable(x)
df.cache()
df.count()
def uncache(spark):
sqlContext = SQLContext(spark.sparkContext)
sqlContext.clearCache()
# df = spark.createDataFrame(customer, s_customer)
# df.registerTempTable('customer')
# print(df.first())
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument("-p", "--prefix")
parser.add_argument("-l", "--load", default='')
parser.add_argument("-ul", "--unload", default=0, type=int)
args = parser.parse_args()
init_spark()
if args.load:
cache(spark, p.tables_need[args.load], args.prefix)
elif args.unload:
uncache(spark)
else:
create_parquet(spark, args.prefix)
# init temp table