Mercurial > hg > MPIWGWeb

annotate wikisqldumpToSQL.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
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
1 import psycopg
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
2
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
3 o = psycopg.connect('dbname=authorities user=dwinter password=3333',serialize=0)
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
4 c = o.cursor()
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
5
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
6 import re
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
7
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
8 def quote(str):
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
9 str=str.replace("'","\\\'")
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
10 str=str.replace("@","\\\'")
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
11 return str#.encode('utf-8')
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
12
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
13 filename="/Users/dwinter/Desktop/dewiki-20060908-categorylinks.sql"
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
14 txt=file(filename).read()
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
15 datas=re.findall("(VALUES.*)",txt)
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
16 for data in datas:
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
17 print data
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
18 txt=data.replace("VALUES","")
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
19 data=re.findall("\((.*?)\)",txt)
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
20 print len(data)
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
21 for result in data:
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
22 result=result.replace("\\\'","@")
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
23
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
24 new=re.findall("\'(.*?)\'",result)
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
25 if len(new)>1:
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
26 if (len(new)>0) and ((new[0]=='Mann') or (new[0]=="Frau")):
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
27 try:
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
28 firstname=new[1].split(",")[1].lstrip()
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
29 except:
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
30 firstname=""
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
31 lastname=new[1].split(",")[0].lstrip()
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
32
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
33 print "INSERT into persons (firstname,lastname) VALUES ('%s','%s')"% (quote(firstname),quote(lastname))
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
34 c.execute("INSERT into persons (firstname,lastname) VALUES ('%s','%s')"% (quote(firstname),quote(lastname)))
bca61e893fcc first checkin of MPIWGWeb r2 branch from CVS into mercurial
casties
parents:
diff changeset
35 c.commit()