is it possible to ignore error and coninue reading from file

[huashan]

I am trying to read from several huge files which may contain embedded null strings in some lines. fread() always stops with message "Jump %d did not finish counting rows exactly where jump %d found its first good line start".

Since I have some many files to read and I don't care to lose some broken lines of data, is it possible to let fread() just ignore the error and continue reading?

[MichaelChirico]

please include information about your version of data.table. if you haven't installed the most recent development version, please do so -- fread has undergone substantial improvements and this concern may be moot.

…

On Nov 22, 2017 11:43 PM, "huashan" ***@***.***> wrote: I am trying to read from several huge files which may contain embedded null strings in some lines. fread() always stops with message "Jump %d did not finish counting rows exactly where jump %d found its first good line start". Since I have some many files to read and I don't care to lose some broken lines of data, is it possible to let fread() just ignore the error and continue reading? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#2496>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AHQQdbwmZWp3OF6Z8m-pGqIjrLaPMuIjks5s5EEwgaJpZM4Qnmf0> .

[huashan]

@MichaelChirico

The data.table version is 1.10.5 release, with the above message "Jump %d did not finish counting rows exactly where jump %d found its first good line start".

I also tried earlier versions of data.table, from 1.9.7 to 1.10.4, all report error message with "embedded null string".

[huashan]

@MichaelChirico

FYI, for version 1.10.5, if I set quote="" in fread() causes R crashes.

[jangorecki]

please provide reproducible example