Michael Malak 2013-02-13, 19:26
Dean Wampler 2013-02-13, 19:47
-Re: INSERT INTO table with STRUCT, SELECT FROM
Edward Capriolo 2013-02-13, 22:49
I developed the inline udtf. Seems to work:
as (id, text) FROM SRC limit 2;
STRUCT (3,'my car?')
) as (id, text) FROM SRC limit 2;
On Wed, Feb 13, 2013 at 2:47 PM, Dean Wampler
<[EMAIL PROTECTED]> wrote:
> Hmm. I tried the following hacks, but all wouldn't parse. Ideas?
> I changed:
> ... select struct(x,y) ...
> ... select struct(x,y) as struct<a:int,b:int> ...
> ... select cast(struct(x,y) as struct<a:int,b:int>) ...
> ... select struct(x as a,y as b) ...
> Okay, but there is a hack that does work; By pass INSERT INTO and just write
> to the directory:
> INSERT DIRECTORY '/path/to/table/directory' SELECT ...;
> Just be careful it doesn't clobber any files already there. I'm paranoid, so
> I would write to a different directory and then move the files over...
> On Wed, Feb 13, 2013 at 1:26 PM, Michael Malak <[EMAIL PROTECTED]>
>> Is it possible to INSERT INTO TABLE t SELECT FROM where t has a column
>> with a STRUCT?
>> Based on
>> I thought perhaps the following would work:
>> echo 1,2 >twovalues.csv
>> CREATE TABLE tc (x INT, y INT) ROW FORMAT DELIMITED FIELDS TERMINATED BY
>> LOAD DATA LOCAL INPATH 'twovalues.csv' INTO TABLE tc;
>> CREATE TABLE oc (z STRUCT<a: int, b: int>);
>> INSERT INTO TABLE oc SELECT struct(x,y) FROM tc;
>> but when I do the above I get:
>> FAILED: SemanticException [Error 10044]: Line 1:18 Cannot insert into
>> target table because column number/types are different 'oc': Cannot convert
>> column 0 from struct<col1:int,col2:int> to struct<a:int,b:int>.
> Dean Wampler, Ph.D.
Michael Malak 2013-02-13, 20:09