Hey,
I just tried out something. Not sure if there is a better way than this.
See if this works for you:
Assuming the text file contains the following line:
$cat /tmp/drill/abc.csv
aaaaabbbbbbbbbb421.5cc
(field1=aaaaa [chars 1-5], field2=bbbbbbbbbb [chars 6-15], field3=42
[digits 16-17], field4=1.5 [digits 18-20], ... )
A query similar to the one shown below can work:
> select
cast(columns[0] as char(5)),
`right`(cast(columns[0] as char(15)),10),
cast(`right`(cast(columns[0] as char(17)),2) as int),
cast(`right`(cast(columns[0] as char(20)),3) as double)
from dfs.tmp.`/drill/abc.csv`;
+------------+------------+------------+------------+
| EXPR$0 | EXPR$1 | EXPR$2 | EXPR$3 |
+------------+------------+------------+------------+
| aaaaa | bbbbbbbbbb | 42 | 1.5 |
+------------+------------+------------+------------+
Regards,
Abhishek
On Mon, Apr 20, 2015 at 3:53 PM, Yousef Lasi <[email protected]> wrote:
> Does anyone have any suggestions on querying fixed-length data files with
> Drill? These are files that are received from a mainframe source and the
> fields within a row are defined by length. For example, field1 = characters
> 1-12, field2 = characters 13-22 etc.
>
> Thanks
>
