A library for loading and processing PBS Professional accounting records.

The PBS Professional scheduler stores event records for each created job in plain-text files called accounting logs. While these records do not contain all of the information presented in the qstat command, they provide a large subset and are useful for gauging system usage patterns, debugging job issues, and more.

PBS Pro does provide the tracejob command to query the accounting logs on the command line.

This pbsparse library allows for easy loading of PBS Pro accounting records into Python scripts for analysis. All of the difficult work of robustly parsing each accounting record is handled for you.

Simply install from PyPI using pip:

$ python3 -m pip install pbsparse

The main interface to load records is the get_pbs_records function. For example, if you wanted to get all jobs by their start record, you could use the following:

from pbsparse import get_pbs_records

job_starts = get_pbs_records("/pbs/accounting/20250301", type_filter = "S")

This function allows for extensive filtering options.

You can also use the PbsRecord class directly:

from pbsparse import PbsRecord

with open("/pbs/accounting/20250301", "r") as pbs_file:
    for line in pbs_file:
        # Read data into object and process metadata
        record = PbsRecord(record, True)
        print(record)