[ 
https://issues.apache.org/jira/browse/ASTERIXDB-1470?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15322079#comment-15322079
 ] 

Abdullah Alamoudi commented on ASTERIXDB-1470:
----------------------------------------------

So I did some reading on this. The 3 octal characters at the beginning 
translates to opening quotes and the 3 octals at the end translates to a 
closing quote. 
In order to read these and convert them correctly, it will take some time. 

So we have 3 options:
1. leave them and read the string as: \342\200\234-1 0.1 0.1 0.5 2e-07 0.001 10 
-1\342\200\235
2. escape them correctly and read the string as: "-1 0.1 0.1 0.5 2e-07 0.001 10 
-1"
3. remove the octal? the string will be read as: -1 0.1 0.1 0.5 2e-07 0.001 10 
-1

Option 1 is the easiest but it doesn't look nice.
Option 2 is the right one but it will take some time to implement.
Option 3 is pretty easy to do but we have to be sure that the octal characters 
will always be something of low significance.

Thoughts?

> large amount of failed entries in failed record log
> ---------------------------------------------------
>
>                 Key: ASTERIXDB-1470
>                 URL: https://issues.apache.org/jira/browse/ASTERIXDB-1470
>             Project: Apache AsterixDB
>          Issue Type: Bug
>          Components: Feeds
>            Reporter: Christopher Green
>            Assignee: Abdullah Alamoudi
>         Attachments: sample_failed_record.log
>
>
> Thank you again for your work in this space.  I have pulled down the new code 
> from the repository and built it Tuesday morning.  I understand from the 
> email message that I would have to remove the instance and create it from 
> scratch.  This wasn't an issue as we have a bunch of data stored up.  Once 
> the instance was deleted, I created the table and feeder and connected the 
> feeder.  We're at about 8 million entries this morning and a count of these 
> entries took about 113 seconds. 
> As of this morning, the failed error log is quite large.  I took a cursory 
> look at the failed entries and they don't appear to be very large entries.  
> Could this be an issue with parsing the classad expressions?
> Let me know if there's anything you need. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to