Using Joern to process a large number of files and generate PDG is a slow solution

I am speechless about this problem . In a previous blog post , I introduced how to pass Joern Generate PDG, Especially in code behavior units PDG： How to use Joern Generate Program Dependence Graph And corresponding to the line number in the source code _ Cricket blog -CSDN Blog

But recently Joern When , Find that the speed is getting slower and slower , I really don't know what's going on . Complain Joern I am not alone in my slow speed , For example, here ：

Today is too slow , I took a close look , Found a... Under the folder workspace Folder , Every time a file is processed, all folders under it will be updated （ for example cpg.bin,cpg.bin1 wait ）, And my directory has 10 Ten thousand folders ！ No wonder it's slow , So the solution is simple ： Empty this folder when running , You can even empty this folder while analyzing each file ：Python Method to empty all files in the specified folder _ Cricket blog -CSDN Blog _python Delete all files in the folder

After this modification , Discovery is much faster ……