BoundsError() when using readtable

[tomrod]

I get this error when trying to run DataFrames' readtable function:

---

BoundsError()
while loading In[4], in expression starting on line 1

in findcorruption at /home/ogre/.julia/v0.3/DataFrames/src/dataframe/io.jl:674
in readtable! at /home/ogre/.julia/v0.3/DataFrames/src/dataframe/io.jl:742
in readtable at /home/ogre/.julia/v0.3/DataFrames/src/dataframe/io.jl:823
in readtable at /home/ogre/.julia/v0.3/DataFrames/src/dataframe/io.jl:890

---

Using version:
Version 0.3.0 (2014-08-20 20:43 UTC)
http://julialang.org release
x86_64-linux-gnu

Appears similar to #586

File is CSV, separator = ',', and charset=us-ascii

enca -L None output:

7bit ASCII characters
CRLF line terminators

[tkelman]

May also be related to #604 - does the file have any comments in it? Does anything change if you run dos2unix on the file?

[tomrod]

Nothing changes if I use dos2unix or fromdos commands to convert (though it does get rid of the pesky CRLF characters).

Seeing now if there is a viable workaround by reading in the table using readdlm and feeding into a DataFrame

[tomrod]

Ah. I found the root issue.

Some of my rows had length 96, some had 92. Removing the 92-length rows resulted in a clean read-in (this done post-dos2unix--no idea if this corrected another bug).

[johnmyleswhite]

Just to clarify: was your data corrupt or did readtable fail to read clean data?

[tomrod]

Failed to replace missing data at end of lines with an NA or other missing value. I can send the data if you would like to try a replication.

[johnmyleswhite]

I think I'm good. It looked something like this:

a,b,c
1,2,3
4,5

Rather than like

a,b,c
1,2,3
4,5,

[tomrod]

That is correct.

[johnmyleswhite]

Ok. I think I'll close this then, since the data isn't fully valid. In the future, we'll need to give better diagnostics when we fail to read a line from an input file.