loss data after sort

[sing1ee]

$ wc -l ngram.data; wc -l ngram_sort.data
 10069731 ngram.data
 10067458 ngram_sort.data

after sorting, I lost about 2273 lines? why?

[cowtowncoder]

Hard to say without data, or code being used (to know input processor). Should not be due to duplicate lines (which are fine). Theoretically could have something to do with differences in linefeed detection -- java-merge-sort accepts all three (\n, \r and \r\n; windows and macos use different ones), not sure if wc does.