Hi, Greg,

The seg fault was caused by a strcmp involving a nil pointer.  We have
added guard for that strcmp.

For your use case, you will need to use the following option

ibis -append 1_row to 4_rows

This tells ibis to append the content in directory 1_row to the
content in directory 4_rows.
Note that the keyword "to" could be neglected.

The updated FastBit code is in SVN revision 536.  Please give it a try
and let us know if you found anything.

Thanks.

John


On 8/7/12 2:22 PM, Greg Barker wrote:
> John -
> 
> Thank you for the updated code. That seems to solve another issue we
> saw where 0 rows being returned when there should have been some rows
> returned. Still need to test the scenario where extra rows were being
> returned.
> 
> I am hitting a segmentation fault when trying to append data (using
> rev 533). The first two attempts to append work fine, but the third is
> failing. I was able to reproduce the issue using ardea, please see
> example below.
> 
> greg@greg-ubuntu:~/fastbit_test$ cat 1_row.csv
> #category_col:k,long_col:l,double_col:d
> AAA,5,35.18
> 
> greg@greg-ubuntu:~/fastbit_test$ cat 4_rows.csv
> #category_col:k,long_col:l,double_col:d
> AAA,1,54.12
> AAA,2,15.13
> AAA,3,51.19
> BBB,4,59.95
> 
> greg@greg-ubuntu:~/fastbit_test$ ardea -d 1_row -m
> "category_col:k,long_col:l,double_col:d" -t 1_row.csv
> ardea: verbose level 0
> Will attempt to parse 1 CSV file
>     1_row.csv
>  with the following column names and types
>     category_col:k,long_col:l,double_col:d
> ardea to read CSV file 1_row.csv ...
> ardea read 1 row from 1_row.csv
> 
> greg@greg-ubuntu:~/fastbit_test$ ardea -d 4_rows -m
> "category_col:k,long_col:l,double_col:d" -t 4_rows.csv
> ardea: verbose level 0
> Will attempt to parse 1 CSV file
>     4_rows.csv
>  with the following column names and types
>     category_col:k,long_col:l,double_col:d
> ardea to read CSV file 4_rows.csv ...
> ardea read 4 rows from 4_rows.csv
> 
> greg@greg-ubuntu:~/fastbit_test$ ibis -d 4_rows -a 1_row
> part[G2DD4CF1A]::append -- committed to use the updated dataset with 1
> rows and 3 columns
> doAppend(1_row): adding 1 row
> 
> greg@greg-ubuntu:~/fastbit_test$ ibis -d 4_rows -a 1_row
> part[G2DD4CF1A]::append -- committed to use the updated dataset with 2
> rows and 3 columns
> doAppend(1_row): adding 1 row
> 
> greg@greg-ubuntu:~/fastbit_test$ ibis -d 4_rows -a 1_row
> Segmentation fault (core dumped)
> 
> Thanks,
> Greg
> 
> On Mon, Aug 6, 2012 at 10:55 AM, K. John Wu <[email protected]
> <mailto:[email protected]>> wrote:
> 
>     Hi, Greg,
> 
>     I did find a bug with the code that handles the string searches.  The
>     fix is checked in as SVN revision 533.  Please give it a try when you
>     get the chance.
> 
>     However, I am not sure it actually addresses the problem you are
>     facing.  If you continue to have problems with your string queries,
>     please let us know.  Please consider giving us a small test case for
>     debugging.
> 
>     Thanks.
> 
>     John
> 
> 
> 
> 
>     On 8/4/12 6:04 PM, Greg Barker wrote:
>     > Hello -
>     >
>     > Similar to the query I mentioned last time, it is of the form:
>     >
>     > select col_a where col_a in ('value1', 'value2')
>     >
>     > col_a in this case is of type text. The problem I'm having is
>     that if
>     > the in clause has more than 1 value in it, every row in the table is
>     > returned, rather than just the rows that match the two values
>     > specified in the in clause. Do you know what might be causing that?
>     >
>     > Thanks,
>     > Greg
> 
> 
_______________________________________________
FastBit-users mailing list
[email protected]
https://hpcrdm.lbl.gov/cgi-bin/mailman/listinfo/fastbit-users

Reply via email to