1
|
#!/usr/bin/env python
|
2
|
# Loads a command's CSV output stream into a PostgreSQL table.
|
3
|
# When no command is specified, just cleans up the specified table.
|
4
|
# The command may be run more than once.
|
5
|
|
6
|
import os.path
|
7
|
import subprocess
|
8
|
import sys
|
9
|
|
10
|
sys.path.append(os.path.dirname(__file__)+"/../lib")
|
11
|
|
12
|
import exc
|
13
|
import opts
|
14
|
import sql
|
15
|
import sql_io
|
16
|
import sql_gen
|
17
|
import streams
|
18
|
import strings
|
19
|
import util
|
20
|
|
21
|
def main():
|
22
|
# Usage
|
23
|
env_names = []
|
24
|
def usage_err():
|
25
|
raise SystemExit('Usage: '+opts.env_usage(env_names)+' '+sys.argv[0]
|
26
|
+' [input_cmd args...]')
|
27
|
|
28
|
# Parse args
|
29
|
input_cmd = sys.argv[1:]
|
30
|
|
31
|
# Get config from env vars
|
32
|
table = opts.get_env_var('table', None, env_names)
|
33
|
schema = opts.get_env_var('schema', 'public', env_names)
|
34
|
has_row_num = opts.env_flag('has_row_num', True, env_names)
|
35
|
db_config = opts.get_env_vars(sql.db_config_names, None, env_names)
|
36
|
verbosity = util.cast(float, opts.get_env_var('verbosity', 3, env_names))
|
37
|
|
38
|
if not (table != None and 'engine' in db_config): usage_err()
|
39
|
|
40
|
# Connect to DB
|
41
|
def log(msg, level=1):
|
42
|
'''Higher level -> more verbose'''
|
43
|
if level <= verbosity:
|
44
|
sys.stderr.write(strings.to_raw_str(msg.rstrip('\n')+'\n'))
|
45
|
db = sql.connect(db_config, log_debug=log)
|
46
|
|
47
|
table = sql_gen.Table(table, schema)
|
48
|
|
49
|
use_copy_from = True
|
50
|
|
51
|
# Loads data into the table using the currently-selected approach.
|
52
|
def load():
|
53
|
# Open input stream
|
54
|
proc = subprocess.Popen(input_cmd, stdout=subprocess.PIPE, bufsize=-1)
|
55
|
in_ = proc.stdout
|
56
|
line_in = streams.ProgressInputStream(in_, sys.stderr, n=1000)
|
57
|
|
58
|
# Import data
|
59
|
try: sql_io.import_csv(db, table, line_in, use_copy_from, has_row_num)
|
60
|
finally:
|
61
|
line_in.close() # also closes proc.stdout
|
62
|
proc.wait()
|
63
|
|
64
|
if input_cmd != []:
|
65
|
try: load()
|
66
|
except sql.DatabaseErrors, e:
|
67
|
if use_copy_from: # first try
|
68
|
exc.print_ex(e, plain=True)
|
69
|
use_copy_from = False
|
70
|
load() # try again with different approach
|
71
|
else: raise
|
72
|
else: sql_io.cleanup_table(db, table)
|
73
|
|
74
|
main()
|