line duplication using logging to file

[seb]

Hi,

Thanks for the help.
Meanwhile I have written the logging function from scratch and it works
without the multiple lines.

This means that the multiple line copy is not due to the multiple
processes (or thread) trying to access the log file but to something
else.

Thanks.
Sebastien.

the new function :


import logging
import time

def write_log(level, message):
	nom_logger="main_log_file.txt"
	logging.basicConfig(level=logging.DEBUG,
                    format='%(asctime)s %(levelname)s %(message)s',
                    filename=nom_logger,
                    filemode='w')

	if str(level).lower() == "info" :
		logging.info(str(message))
	elif str(level).lower() =="error":
		logging.error(str(message))
	elif str(level).lower()=="warning" :
		logging.warning(str(message))
	elif str(level).lower() =="critical":
		logging.critical(str(message))
	elif str(level).lower() == "exception":
		logging.exception(str(message))
	else :
		logging.INFO(str(message))

	return







Paul McGuire a écrit :
> "seb" <sebastien.thur at laposte.net> wrote in message
> news:1168329975.491514.282000 at 11g2000cwr.googlegroups.com...
> > Hi, I am writing to a file some basic information using the logging
> > module.  It is working but in the log file some line are printed
> > several time. I had put some print debugging messages in the logging
> > function (so they appear on the consile) and they are called once only.
> > Obviously there is some understantding of the logging module that I am
> > missing.
> >
> > My simple logging program (see below) is called by several processes.
> > In this way I can collect the information from various sources (and not
> > use the network enabled logging module)
> >
> > I am using python 2.4 on WinXP SP2.
> >
> > Do you have any idea ? Thanks in advance.
> >
> > Seb.
> >
>
> A quick tally of log messages by timestamp and comment gives this data:
>
> ('2007-01-08 18:26:19,578', '___init_rs232initrs232_openCOM1') : 1
> ('2007-01-08 18:26:19,578', '___run____thread lance') : 2
> ('2007-01-08 18:26:32,015', '___test1TEST 1 = OK') : 3
> ('2007-01-08 18:26:42,483', '___test1TEST 1 = OK') : 4
> ('2007-01-08 18:26:53,750', '___test1TEST 1 = OK') : 5
> ('2007-01-08 18:27:03,092', '___test1TEST 1 = OK') : 6
> ('2007-01-08 18:27:13,671', '___test1TEST 1 = OK') : 7
> ('2007-01-08 18:27:14,796', '___run___fin dans le run car continue = 0') : 8
> ('2007-01-08 18:27:14,890', "___stopthread demande d'arret") : 9
> ('2007-01-09 08:51:26,562', '___init_rs232initrs232_openCOM1') : 1
> ('2007-01-09 08:51:26,733', '___run____thread lance') : 2
> ('2007-01-09 08:51:39,453', '___test1TEST 1 = OK') : 3
> ('2007-01-09 08:51:48,280', '___test1TEST 1 = OK') : 4
> ('2007-01-09 08:51:58,750', '___test1TEST 1 = OK') : 5
> ('2007-01-09 08:52:09,812', '___test1TEST 1 = OK') : 6
> ('2007-01-09 08:52:19,078', '___test1TEST 1 = OK') : 7
> ('2007-01-09 08:52:22,078', '___run___fin dans le run car continue = 0') : 8
> ('2007-01-09 08:52:22,125', "___stopthread demande d'arret") : 8
> ('2007-01-09 08:52:22,125', "___stopthread demande d'arret ") : 1
>
> Does this suggest anything to you?
>
> -- Paul
>
>
> (BTW, here is the pyparsing program I used to do this analysis)
>
> from pyparsing import
> Word,nums,Combine,alphas,oneOf,Literal,SkipTo,restOfLine
>
> # create pyparsing grammar definition for a log line
> date = Word(nums,exact=4)+'-'+Word(nums,exact=2)+'-'+Word(nums,exact=2)
> time = Word(nums,exact=2)+':'+Word(nums,exact=2)+':'+Word(nums,exact=2)+ \
>         ','+Word(nums,exact=3)
> timestamp = Combine(date + ' ' + time)
> severity = oneOf( ["INFO","WARNING"] ) # not complete, but enough for this
> data
> backslash = Literal("\\")
> fileref = Combine(Word(alphas,exact=1)+":" + backslash + SkipTo(".py") +
> ".py")
> logline = ( timestamp.setResultsName("timestamp") + "-" +
>             severity.setResultsName("severity") + "-" +
>             fileref.setResultsName("fileref") +
>             restOfLine.setResultsName("comment") )
>
> # create list of ParseResults, with addressable log line elements by results
> name
> logEntries = [ logline.parseString(line) for line in logdata ]
>
> # tally up log lines by timestamp and comment
> tallyByTimestamp = {}
> for entry in logEntries:
>     tallyKey = (entry.timestamp, entry.comment)
>     tallyByTimestamp[tallyKey] = tallyByTimestamp.get(tallyKey,0) + 1
>
> for ts in sorted( tallyByTimestamp.items() ):
>     print "%s : %d" % ts