1

Closed

Cannot convert value "(PDH-CSV 4.0) (E. Australia Standard Time)(-600)" to type "System.DateTime". Error: "String was not recognized as a valid DateTime."

description

Hello all,
 
I am having a problem when I try to run PAL against CSV files that contain date time as:
 
"04/10/2012 14:12:45.244","33.089195026026793","41.443193273460409","15.417207996535486"
 
It appears that the date entry causes the issue. Similar CSV files that do not have the date range of 04/01/2012 - 04/12/2012 are fine. It seems PAL thinks the date range is for 04th january through 04th December, rather than 1st April through to 12th of April.
 
This issue is occurring on a number of servers at different sites.
 
I am uploading two CSV files produced in a similar fashion, but covering different dates. One works and the other fails when parsing through PAL.
 
Any help on modifying the date column is greatly appreciated.
 
Cheers,
Henry

file attachments

Closed Jun 13, 2012 at 5:02 AM by ClintH
The user edited the CSV file - potential corruption in the CSV file.

comments

hg0mez wrote Jun 12, 2012 at 11:15 PM

In addition, there are two files here, one runs successfully to completion when using PAL:

health_check_bnesrv004_2012041612 - this file is ok to parse.csv

And the other file "ausgemilt001.csv" fails with the error message:

"Cannot convert value "(PDH-CSV 4.0) (E. Australia Standard Time)(-600)" to type "System.DateTime". Error: "String was not recognized as a valid DateTime."


If someone can help me modify the first column (i.e date time field) or the actual TIMEZONE declaration at the top, this would be great. The timezone is in fact Eastern Australian Time Zone, but the date time values appear as MM/DD/YY HH:MM:SS - which may lead to some confusion? All servers are configured in the same way. Some of the CSV files run to completion and others fail. I even tried running a relog.exe and modifying all CSV files to Binary but the relog process also failed when I try to run it for multiple files.

hg0mez wrote Jun 13, 2012 at 12:11 AM

Guys and Gals,


After a bit of head scratching I managed to work out the issue here. As I ran a copy *.csv new_file.csv - all of the headers from each csv was copied into the final file and the parser complained that it was trying to parse a line that had "(PDH-CSV 4.0) (E. Australia Standard Time)(-600)" in the first column, first row.

I deleted each of these lines from the resulting csv and it worked a treat.

This issue can be closed.

Cheers,
Henry

ClintH wrote Jun 13, 2012 at 5:01 AM

The first row is used to define the column names and PAL is aware and is expecting this format. I'm not sure why your change actually worked though. There seems to be a lot of people using CSV as counter logs. I highly encourage you to use binary (.blg) counter logs because they track all of the counter instance that come and go while CSV counter logs will not track any new counter instances after you start the data collector. Always use binary (.blg) data collectors! In any case, I'm happy you were able to resolve the issue on your own and thank you for posting your solution.