search

logpai / loghub-2.0

A Large-scale Evaluation for Log Parsing Techniques: How Far are We? [ISSTA'24]
Other
51 stars 7 forks source link

BGL template #4

Closed sleep-zzw-bot closed 3 months ago

sleep-zzw-bot commented 3 months ago

Why do all the structured generated by BGL logs have no label attribute? Will this affect subsequent use? I hope you can tell me where the structured with label is?

zhjiang22 commented 3 months ago

Hi, thank you for your interest in our work. The "label" column, as well as other columns such as "node" and "component," are extracted using pre-defined regular expressions as part of the preprocessing step. In the task of log parsing, our main focus is on parsing the content of log messages. To reduce the size of the generated structured CSV file, we have removed these unrelated columns and retained only the following columns: LineId, Content, EventId, EventTemplate.

If you need to access the deleted columns, you can utilize the regular expressions in the benchmark_settings of the logparser/logpai repository to process the original log file. Hope this could be helpful for you. :blush: