Mercurial > hg > MPIWGWeb

comparison wikixml2sql.py @ 0:bca61e893fcc

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
first checkin of MPIWGWeb r2 branch from CVS into mercurial
author casties
date Thu, 10 Jan 2013 17:52:13 +0100
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:bca61e893fcc
1 import xml.parsers.expat
2 import psycopg
3
4 filename="mann.xml"
5
6 # 3 handler functions
7 global toggle
8 toggle=False
9
10 global c
11
12 def quote(str):
13 str=str.replace("'","\\\'")
14 return str.encode('utf-8')
15
16 def start_element(name, attrs):
17 global toggle
18 if name=="title":
19 toggle=True
20 def end_element(name):
21 global toggle
22 if name=="title":
23 toggle=False
24 def char_data(data):
25 global toggle
26 global c
27 if toggle:
28 splitted=data.split()
29 if splitted >1:
30 lastname=splitted[-1]
31 firstname=" ".join(splitted[0:-1])
32 else:
33 lastname=splitted[0]
34 firstname=""
35 print "INSERT into persons (firstname,lastname) VALUES ('%s','%s')"% (quote(firstname),quote(lastname))
36 c.execute("INSERT into persons (firstname,lastname) VALUES ('%s','%s')"% (quote(firstname),quote(lastname)))
37 c.commit()
38
39
40 o = psycopg.connect('dbname=authorities user=dwinter password=3333',serialize=0)
41 c = o.cursor()
42
43 p = xml.parsers.expat.ParserCreate()
44
45 p.StartElementHandler = start_element
46 p.EndElementHandler = end_element
47 p.CharacterDataHandler = char_data
48 fh=file(filename)
49 p.ParseFile(fh)
50
51 o.close()