I got this working using the following example:
*Data:*
$ cat abc.ssv
"a b" c
*Format plugin:*
"ssv": {
"type": "text",
"extensions": [
"ssv"
],
"delimiter": " "
}
*Query:*
> select * from dfs.tmp.`abc.ssv`;
+--------------+
| columns |
+--------------+
| ["a b","c"] |
+--------------+
1 row selected (0.45 seconds)
On Wed, Oct 14, 2015 at 12:16 PM, Sungwook Yoon <[email protected]> wrote:
> Then, I think the storage plugin is not functioning correctly.
>
> I am testing with drill 1.2 on mapr 5.0
>
> Sungwook
>
>
> On Wed, Oct 14, 2015 at 1:53 PM, Jacques Nadeau <[email protected]>
> wrote:
>
> > Abhishek is correct, we currently only display and store non-default
> values
> > for the text format configuration options (including quote). The defaults
> > are here:
> >
> >
> >
> https://github.com/apache/drill/blob/master/exec/java-exec/src/main/java/org/apache/drill/exec/store/easy/text/TextFormatPlugin.java#L137
> >
> >
> > --
> > Jacques Nadeau
> > CTO and Co-Founder, Dremio
> >
> > On Wed, Oct 14, 2015 at 9:46 AM, Abhishek Girish <
> > [email protected]>
> > wrote:
> >
> > > I was able to provide comma (",") single quote ("'"), Unicode (\u0000
> > > \u0001 .. \u0006) and a few other single characters - the quote field
> did
> > > show up after updating.
> > >
> > > Providing "\"" made the quote field disappear after update. It could be
> > > because double quote or "\u0022" could be the default.
> > >
> > > Also, giving Unicode characters such as "\u0009", "\u0023" updated the
> > > value to corresponding ASCII types ("\t", "#", ...).
> > >
> > > When you say Drill fails to read data correctly, can you give some
> > > specifics? I'd expect it to work.
> > >
> > > -Abhishek
> > >
> > > On Tue, Oct 13, 2015 at 2:54 PM, Sungwook Yoon <[email protected]>
> > wrote:
> > >
> > > > Andries,
> > > >
> > > > I put quote before the delimiter declaration, did not change the
> > > behavior.
> > > > That is, the "quote" field disappears after successful update.
> > > >
> > > > I tried "\u0000".
> > > > With this, storage plugin updates successfully and changes the
> storage
> > > > plugin definition.
> > > >
> > > > But, of course, this is not helping the situation in terms of
> querying.
> > > >
> > > > Thanks,
> > > >
> > > > Sungwook
> > > >
> > > >
> > > > On Tue, Oct 13, 2015 at 4:38 PM, Andries Engelbrecht <
> > > > [email protected]> wrote:
> > > >
> > > > > Does the quote field in the SP config disappear after you click
> > update?
> > > > > Did you check if it works if you put it before delimiter?
> > > > >
> > > > > Have you tried "quote": "\u0000"?
> > > > >
> > > > >
> > > > > —Andries
> > > > >
> > > > >
> > > > > > On Oct 13, 2015, at 1:24 PM, Sungwook Yoon <[email protected]>
> > > wrote:
> > > > > >
> > > > > > I have an apache access log.
> > > > > > It is " " separated '"' quoted string per line.
> > > > > >
> > > > > > In the storage plugin, I defined the storage plugin as,
> > > > > > "ssv": {
> > > > > > "type": "text",
> > > > > > "extensions": [
> > > > > > "ssv"
> > > > > > ],
> > > > > > "delimiter": " ",
> > > > > > "quote": "\""
> > > > > > },
> > > > > >
> > > > > > The "quote" field, is accepted but after update on the storage
> > > plugin,
> > > > > with
> > > > > > success message, "quote" field disappears.
> > > > > >
> > > > > > WHY?
> > > > > >
> > > > > > I tried to define the "quote" as "\u0022". Again disappears with
> > > > success
> > > > > > message.
> > > > > >
> > > > > > Then, drill fails to read the data correctly, ignoring the
> "quoted"
> > > > value
> > > > > > direction.
> > > > > >
> > > > > > What should I do in this case?
> > > > > >
> > > > > > Thanks,
> > > > > >
> > > > > > Sungwook
> > > > >
> > > > >
> > > >
> > >
> >
>
