Thanks for the help. The ravel in front of the character I wanted to
test with (=< , '1') was just what I needed. So now I can perform a
readcsv to pull in my data and perform the correct test and then apply
the answer to the original table to pull out only the records I am
interested in with the copy operator (dyadic form of #).
Thanks again to Rob and Brian for their knowledge
Tom McGuire
Rob Hodgkinson wrote:
> In <mydata> each box contains a 1 element vector...
> $each mydata
> +-+-+-+-+-+-+-+-+-+-+
> |1|1|1|1|1|1|1|1|1|1|
> +-+-+-+-+-+-+-+-+-+-+
>
> In <lin> each box contains a scalar (atom with no shape)
> $each lin
> +++++++++++
> |||||||||||
> +++++++++++
>
> When you test each variable against the test item of <'1',
> the shape doesn't match in mydata, but does in lin:
> mydata = <'1'
> 0 0 0 0 0 0 0 0 0 0
> lin = <'1'
> 0 1 0 0 0 0 0 0 0 0
>
> Likewise if you test against a test item of <,'1' vice versa:
> mydata = <,'1'
> 0 1 0 0 0 0 0 0 0 0
> lin = <,'1'
> 0 0 0 0 0 0 0 0 0 0
>
> You can also apply the verb '=' within each box, so try this ...
> ]mydata1 =: 'test1';(2 3$'1234');'not 1 to test'
> +-----+---+-------------+
> |test1|123|not 1 to test|
> | |412| |
> +-----+---+-------------+
> '1' =&.> mydata1 NB. This tests for '1' inside each box ...
> +---------+-----+-------------------------+
> |0 0 0 0 1|1 0 0|0 0 0 0 1 0 0 0 0 0 0 0 0|
> | |0 1 0| |
> +---------+-----+-------------------------+
>
> Note =&.> can use the cover name 'each' (for &.>) as follows:
> '1' = each mydata1
> +---------+-----+-------------------------+
> |0 0 0 0 1|1 0 0|0 0 0 0 1 0 0 0 0 0 0 0 0|
> | |0 1 0| |
> +---------+-----+-------------------------+
>
> When you use {. you need to apply it within each box:
> ({. each mydata)=<'1'
> 0 1 0 0 0 0 0 0 0 0'
>
> Hope this helps.../Rob Hodgkinson
>
> On 27/02/08 4:21 PM, "The World" <[EMAIL PROTECTED]> wrote:
>
>
>> I've been playing with the readcvs function and am having a problem with
>> the boxed data. I can read in a 2 dimensional array of boxed character
>> data. I can select any particular column using
>>
>> col =. idx {"_1 myboxeddata
>>
>> The problem comes in when I want to test each of the boxes in the column
>> to see which one is equal to the single number 1 (as a character of
>> course). I would like to generate an array of 1s and 0s, the 1s being
>> where the number 1 character is. I would like to do this in the
>> following fashion
>>
>> col = <'1'
>>
>> But I always get 0s as the answer even though the boxed list contains a
>> single boxed 1 character
>>
>> I've recreated the problem in simpler form:
>>
>> mydata =. |: ;: 1 ]\'0123456789'
>> mydata
>> │0│1│2│3│4│5│6│7│8│9│
>>
>> mydata = <'1'
>> 0 0 0 0 0 0 0 0 0 0
>>
>> $mydata
>> 1 10
>>
>> So because this horizontal format of mydata was transposed from a column
>> the shape is a little different than if I built this directly using ;
>>
>> lin =. '0';'1';'2';'3';'4';'5';'6';'7';'8';'9'
>> lin
>> │0│1│2│3│4│5│6│7│8│9│
>> $lin
>> 10
>> lin = <'1'
>> 0 1 0 0 0 0 0 0 0 0
>>
>> which works the way I suspect. Even if I use take ({.) to get the
>> dimensions of the boxed list mydata the same as lin in the previous
>> example I still get all 0s:
>>
>> {. mydata
>> │0│1│2│3│4│5│6│7│8│9│
>> ${. mydata
>> 10
>> ({. mydata) = <'1'
>> 0 0 0 0 0 0 0 0 0 0
>>
>> So what is going on here that I am missing?
>>
>>
>>
>>
>>
>> ----------------------------------------------------------------------
>> For information about J forums see http://www.jsoftware.com/forums.htm
>>
>
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
