cProfile, how to understand why lambda function is most time consuming call

Question

I'm new to cprofile but I'm running it on a python tool written for ArcGIS. It tells me that the most time consuming function is the lambda function in _base.py:

ncalls	tottime	percall	cumtime	percall	filename:lineno(function)
17875	137.877	0.008	200.404	0.011	_base.py:510(<lambda>)
962732	67.568	0	67.568	0	{built-in method createobject}
7497	44.223	0.006	44.223	0.006	{built-in method distanceto}
8301909/8300996	35.652	0	35.659	0	{getattr}
713	35.235	0.049	35.235	0.049	{built-in method listfields}

_base.py is an arcpy file in this location:

C:\Program Files (x86)\ArcGIS\Desktop10.8\arcpy\arcpy\geoprocessing

How can I dig deeper and find out what's using this lambda function?

The tool is over 5000 lines of code so I can't post it but I'm importing cprofile at the start:

    import cProfile, pstats
    profiler = cProfile.Profile()
    profiler.enable()

then disabling it and dumping the stats just before the end:

    profiler.disable()
    stats.dump_stats(myStatsLocation)

Answer 1

Usually when I need to debug slow code, I use line profiler. Using this you can use decorator on suspect slow functions. From there, you can find out, like with cProfiler, what's being called and how many times.

The cProfiler already pinpoints to line 510 in your _base.py. So probably that lambda calls another function which is slow. Adding the @profile on that function might shed some light on your performance issues.