When parsing CSV files that require a custom date format, the performance degrades a lot. By not creating a new DateFormat for every line in the file, the performance can be improved a lot (with minimal change to current code base). In my own code I saw up to 50x improvements (but I guess that depends on the exact date format).
When parsing CSV files that require a custom date format, the performance degrades a lot. By not creating a new DateFormat for every line in the file, the performance can be improved a lot (with minimal change to current code base). In my own code I saw up to 50x improvements (but I guess that depends on the exact date format).
I created a pull request for this.