I have 25GB TSV file and trying to import it with command:
D:\sphinx\bin>indexer.exe -c D:\sphinx\sphinx.conf products --rotate
It works some time, but then shows error
ERROR: index 'products': source 'products_tsv': read error 'No error' (line=4595827, pos=908, docid=4595827).
But record at line 4595827 have no problems. I have two questions:
Lost a lot of time on checking datafile and found a lot of hidden symbols such as SYM (\U001A
), NULL (\0000
) and a more of them, which turns Sphinx crazy.
Simply(if «simply» can be said about 25GB file) replaced all SYM to ' and removed others.
We moved forward and faced another issue, but this is another question.