search

pjfanning / poi-shared-strings

Memory efficient Shared Strings Table implementation for POI streaming
Apache License 2.0
8 stars 7 forks source link

for fullFormat=false store plain strings in map #25

Closed pjfanning closed 2 years ago

pjfanning commented 2 years ago

https://github.com/pjfanning/poi-shared-strings/issues/24

pjfanning commented 2 years ago

@mirahbo I still need to fix a broken test and add extra test coverage - when I get this working, I'll look at the temp file comments table (which you are probably not using)

pjfanning commented 2 years ago

@mirahbo this is merged - the perf definitely looks much better with the changes (but only when fullFormat=false) - I can get a poi-shared-strings release done in next few days (and plugin the update to excel-streaming-reader) - if you think this change is enough, I can press on with those releases - hopefully, tomorrow

mirahbo commented 2 years ago

@pjfanning I'm going to run a profiling with this new version. I think that it will meet our own requirements, because as you said, we need neither reading the comments table nor using fullFormat. But many other users will enjoy the performance boost, I guess.

mirahbo commented 2 years ago

@pjfanning I did the test within our own application. I can't believe how much faster this is 🤣 The whole processing takes 20s and reading the 1 million values is not a CPU hot spot anymore 🎉