| Re: edit a large dataset [message #56091] |
Wed, 03 October 2007 11:54 |
queiny
Messages: 15
Registered: November 2005
|
Junior Member |
|
|
Thanks. I don't know we can use
data_orig[a] = data_sorted.
that is the inversion of sort function.
We don't do landuse change modeling. The data we are using provide
landuse infor for reference.
On Oct 3, 12:29 pm, Jean H <jghas...@DELTHIS.ucalgary.ANDTHIS.ca>
wrote:
>
> well, you just have to do
> data_orig[a] = data_sorted
> or similarly
> newLandUse[a] = new_sorted_land_use
>
> Jean
>
> PS: are you doing some kind of land-use change model?
|
|
|
|
| Re: edit a large dataset [message #56092 is a reply to message #56091] |
Wed, 03 October 2007 11:43  |
queiny
Messages: 15
Registered: November 2005
|
Junior Member |
|
|
yeah, sort(a) will bring it back to original order. Thanks!!
>
> a=sort(data_orig)
> data_sorted=data_orig[a]
> data_orig=data_sorted[sort(a)]
>
> Cheers,
> Vince
|
|
|
|
| Re: edit a large dataset [message #56095 is a reply to message #56092] |
Wed, 03 October 2007 09:29 |
Jean H.
Messages: 472
Registered: July 2006
|
Senior Member |
|
|
> a=sort(data_orig)
> data_sorted=data_orig(a)
> ;processing based on 'data_sorted'
>
> but how could I go back the order of 'data_orig'? Is there a inversion
> of 'sort'.
>
> Thanks,
Hi,
well, you just have to do
data_orig[a] = data_sorted
or similarly
newLandUse[a] = new_sorted_land_use
Jean
PS: are you doing some kind of land-use change model?
|
|
|
|
| Re: edit a large dataset [message #56096 is a reply to message #56095] |
Wed, 03 October 2007 09:45 |
Vince Hradil
Messages: 574
Registered: December 1999
|
Senior Member |
|
|
On Oct 3, 11:40 am, hradilv <hrad...@yahoo.com> wrote:
> On Oct 3, 9:48 am, queiny <quein...@yahoo.com> wrote:
>
>
>
>> Hi, I have a large dataset (>1e7 records). One field of the record is
>> a string 'landuse'. Now I want to add one more field based on the
>> 'landuse', i.e., if it is 'forest', then add 'aa', if it is 'urban',
>> then add 'bb', etc.. Since there are more than 20 different landuse,
>> use 'if-then', or 'case' for each record will be quite inefficient.
>
>> I am thinking of 'sort' the records based on 'landuse', and then use
>> 'uniq' to locate the starting point and number of records for each
>> 'landuse' type and add the new fields accordingly. But how could I go
>> back to the original order afterward?
>
>> a=sort(data_orig)
>> data_sorted=data_orig(a)
>> ;processing based on 'data_sorted'
>
>> but how could I go back the order of 'data_orig'? Is there a inversion
>> of 'sort'.
>
>> Thanks,
>
> a=sort(data_orig)
> data_sorted=data_orig[a]
> data_orig=data_sorted[sort(a)]
>
> Cheers,
> Vince
I was considering:
data_orig = data_orig[ sort( data_orig[ sort(data_orig) ] ) ]
But thought it was little to obfuscatory 8)
|
|
|
|
| Re: edit a large dataset [message #56098 is a reply to message #56095] |
Wed, 03 October 2007 09:40 |
Vince Hradil
Messages: 574
Registered: December 1999
|
Senior Member |
|
|
On Oct 3, 9:48 am, queiny <quein...@yahoo.com> wrote:
> Hi, I have a large dataset (>1e7 records). One field of the record is
> a string 'landuse'. Now I want to add one more field based on the
> 'landuse', i.e., if it is 'forest', then add 'aa', if it is 'urban',
> then add 'bb', etc.. Since there are more than 20 different landuse,
> use 'if-then', or 'case' for each record will be quite inefficient.
>
> I am thinking of 'sort' the records based on 'landuse', and then use
> 'uniq' to locate the starting point and number of records for each
> 'landuse' type and add the new fields accordingly. But how could I go
> back to the original order afterward?
>
> a=sort(data_orig)
> data_sorted=data_orig(a)
> ;processing based on 'data_sorted'
>
> but how could I go back the order of 'data_orig'? Is there a inversion
> of 'sort'.
>
> Thanks,
a=sort(data_orig)
data_sorted=data_orig[a]
data_orig=data_sorted[sort(a)]
Cheers,
Vince
|
|
|
|
| Re: edit a large dataset [message #56100 is a reply to message #56095] |
Wed, 03 October 2007 07:58 |
Peter Clinch
Messages: 98
Registered: April 1996
|
Member |
|
|
queiny wrote:
> Hi, I have a large dataset (>1e7 records). One field of the record is
> a string 'landuse'. Now I want to add one more field based on the
> 'landuse', i.e., if it is 'forest', then add 'aa', if it is 'urban',
> then add 'bb', etc.. Since there are more than 20 different landuse,
> use 'if-then', or 'case' for each record will be quite inefficient.
I generally find Perl is much better for hacking data sets than IDL,
because systematically fiddling with file contents is what it does very
well. Once it's been suitably butchered, then into IDL for the actual
crunching of numbers that IDL does very well.
Pete.
--
Peter Clinch Medical Physics IT Officer
Tel 44 1382 660111 ext. 33637 Univ. of Dundee, Ninewells Hospital
Fax 44 1382 640177 Dundee DD1 9SY Scotland UK
net p.j.clinch@dundee.ac.uk http://www.dundee.ac.uk/~pjclinch/
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
