Missing record delimiter "\n", saw EOF instead

rajan.n · Post by **rajan.n** » Wed Oct 18, 2006 8:10 am

Hi thanx a lote for the responce,
according to my job, am having a seq file as i said it clearly ..that am reading a xml file . this reads as a single coloumn. and the xml input stage will parse in to diff coloumns.
so am getting this error in the initial stage ..tht is seq stage .that have only one column.

kumar_s · Post by **kumar_s** » Wed Oct 18, 2006 7:20 pm

Change the Char data type to Varchar.
As character Datatype might expect the fixed size character for the defined length. Since you input seems variable length, there is a chance where new line character may come even before the defined length.

rexsmith · Post by **rexsmith** » Fri Sep 07, 2007 4:28 pm

I've checked all the threads for an answer but i do not see one. I have the same issue . I had one job which was working with no issues couple of days back now i try to view the data it shows the same error. First thing, i was getting this from another job which i was developing today so i just went to check the old job it started throwing same warning . What could be the reason ?

ray.wurlod · Post by **ray.wurlod** » Fri Sep 07, 2007 6:44 pm

What's changed? ("Nothing" is not the correct answer. Dig.)

shiva_reddys447 · Post by **shiva_reddys447** » Sat Sep 08, 2007 12:25 am

if the data is processing to next set of stages properly we can neglect this warning

rexsmith · Post by **rexsmith** » Tue Sep 11, 2007 3:14 pm

I had $APT_IMPEXP_CHARSET set in the job . My source file is a ASCII converted file from mainframe .This variable is not allowing me to read the file . Usage of this variable has an effect on the other job which has source in ASCII .This could be one of the reason but iam not really sure.

asaf_arbely · Post by **asaf_arbely** » Sun Aug 10, 2014 2:25 am

I had the same warning ("Missing record delimiter "", saw EOF instead") while reading from HDFS with BDFS stage.

Changing the "Final delimiter =none" or "=end" didn't help.

When I converted the files from DOS format to UNIX format I stopped getting this error (BDFS stage doesn't have the option to choose DOS/Unix format like in seq file).

vgupta88 · Post by **vgupta88** » Thu Jul 06, 2017 4:01 am

I also faced the same issue. Below 2 points made my job successfully finish without any warnings:

1) My file should follow the below format. Instead of Case2, follow Case1:

Case 1

ABCTTTAAAABBB
ABCTTTAAAABBB
ABCTTTAAAABBB
ABCTTTAAAABBB
|

Case 2

ABCTTTAAAABBB
ABCTTTAAAABBB
ABCTTTAAAABBB
ABCTTTAAAABBB|

2) Since my file was of DOS format, I made my format tab settings look like:
Final Delimiter - none
Record Delimiter String - DOS format

Thanks!