Good afternoon.
In analysis_results table I always see ERROR: no runtime found in stratum_1 column when analysis_id is >= 20000001.
I am running the Achilles version that is currently in main branch (1.8).
Further debugging led me to the conclusion that the issue happens for performance benchmark analyses.
The execution times are searched for in logs and the matches converted to a dataframe. The error is shown when the dataframe's length is not 1. In this case, It will be always 0 since it expects matching on snake_case names but in actuality the log contains them in camelCase (e.g analysis_id vs analysisId).
Some additional considerations:
- My DB engine is PostgreSQL.
- Since the library uses its own logs as part of the input for computing the entries in analysis_results, I always heed the warning of deleting the output directory before each run that I have seen in older bug reports.
- I have forked main branch and comited a fix/workaround that could be contributed or used as help for understanding the problem. It can be seen on : main...davidpedrosa:Achilles:main
Best regards
Good afternoon.
In analysis_results table I always see ERROR: no runtime found in stratum_1 column when analysis_id is >= 20000001.
I am running the Achilles version that is currently in main branch (1.8).
Further debugging led me to the conclusion that the issue happens for performance benchmark analyses.
The execution times are searched for in logs and the matches converted to a dataframe. The error is shown when the dataframe's length is not 1. In this case, It will be always 0 since it expects matching on snake_case names but in actuality the log contains them in camelCase (e.g analysis_id vs analysisId).
Some additional considerations:
Best regards