Python logging - new log file each loop iteration

Question

I would like to generate a new log file on each iteration of a loop in Python using the logging module. I am analysing data in a for loop, where each iteration of the loop contains information on a new object. I would like to generate a log file per object.

I looked at the docs for the logging module and there is capability to change log file on time intervals or when the log file fills up, but I cannot see how to iteratively generate a new log file with a new name. I know ahead of time how many objects are in the loop.

My imagined pseudo code would be:

import logging

for target in targets:
    logfile_name = f"{target}.log"
    logging.basicConfig(format='%(asctime)s - %(levelname)s : %(message)s',
                        datefmt='%Y-%m/%dT%H:%M:%S',
                        filename=logfile_name,
                        level=logging.DEBUG)


    # analyse target infomation
    logging.info('log target info...')

However, the logging information is always appended to the fist log file for target 1.

Is there a way to force a new log file at the beginning of each loop?

Answer 1

This might not be the best solution, but it will create new log file for each iteration. What this is doing is, adding a new file handler in each iteration.

import logging
targets = ["a", "b", "c"]
logger = logging.getLogger(__name__)
logger.setLevel(logging.INFO)

for target in targets:
    log_file = "{}.log".format(target)
    log_format = "|%(levelname)s| : [%(filename)s]--[%(funcName)s] : %(message)s"
    formatter = logging.Formatter(log_format)

    # create file handler and set the formatter
    file_handler = logging.FileHandler(log_file)
    file_handler.setFormatter(formatter)

    # add handler to the logger
    logger.addHandler(file_handler)

    # sample message
    logger.info("Log file: {}".format(target))

Answer 2

This is not necessarily the best answer but worked for my case, and just wanted to put it here for future references. I created a function that looks as follows:

def logger(filename, level=None, format=None):
    """A wrapper to the logging python module

    This module is useful for cases where we need to log in a for loop
    different files. It also will allow more flexibility later on how the
    logging format could evolve.

    Parameters
    ----------
    filename : str
        Name of logfile. 
    level : str, optional
        Level of logging messages, by default 'info'. Supported are: 'info'
        and 'debug'.
    format : str, optional
        Format of logging messages, by default '%(message)s'.

    Returns
    -------
    logger
        A logger object.
    """

    levels = {"info": logging.INFO, "debug": logging.DEBUG}

    if level is None:
        level = levels["info"]
    else:
        level = levels[level.lower()]

    if format is None:
        format = "%(message)s"

    # https://stackoverflow.com/a/12158233/1995261
    for handler in logging.root.handlers[:]:
        logging.root.removeHandler(handler)

    logger = logging.basicConfig(filename=filename, level=level, format=format)

    return logger

As you can see (you might need to scroll down the code above to see the return logger line), I am using logging.basicConfig(). All modules I have in my package that log stuff, have the following at the beginning of the files:

import logging
import other stuff

logger = logging.getLogger()


class SomeClass(object):
    def some_method(self):
        logger.info("Whatever")
        .... stuff

When doing a loop, I have call things this way:

if __name__ == "__main__":
    for i in range(1, 11, 1):
        directory = "_{}".format(i)

        if not os.path.exists(directory):
            os.makedirs(directory)
        filename = directory + "/training.log"
        logger(filename=filename)

I hope this is helpful.

Answer 3

Here is a working version for this problem. I was only able to get it to work if the targets already have .log before going into the loop so you may want to add one more for before going into targets and override all targets with .log extension

import logging

targets = ["a.log","b.log","c.log"]
for target in targets:

    log = logging.getLogger(target)
    formatter = logging.Formatter('%(asctime)s - %(levelname)s : %(message)s', datefmt='%Y-%m/%dT%H:%M:%S')
    fileHandler = logging.FileHandler(target, mode='a')
    fileHandler.setFormatter(formatter)
    streamHandler = logging.StreamHandler()
    streamHandler.setFormatter(formatter)
    log.addHandler(fileHandler)
    log.addHandler(streamHandler)

    log.info('log target info...')

Answer 4

I'd like to slightly modify @0Nicholas's method. The direction is right, but the first FileHandler will continue log information into the first log file as long as the function is running. Therefore, we would want to pop the handler out of the logger's handlers list:

import logging
targets = ["a", "b", "c"]
logger = logging.getLogger(__name__)
logger.setLevel(logging.INFO)
log_format = "|%(levelname)s| : [%(filename)s]--[%(funcName)s] : %(message)s"
formatter = logging.Formatter(log_format)

for target in targets:
    log_file = f"{target}.log"
    
    # create file handler and set the formatter
    file_handler = logging.FileHandler(log_file)
    file_handler.setFormatter(formatter)

    # add handler to the logger
    logger.addHandler(file_handler)

    # sample message
    logger.info(f"Log file: {target}")
    # close the log file
    file_handler.close()
    # remove the handler from the logger. The default behavior is to pop out      
    # the last added one, which is the file_handler we just added in the 
    # beginning of this iteration.
    logger.handlers.pop()