| Re: indexing over structure tags [message #24802] |
Thu, 19 April 2001 12:23  |
tam
Messages: 48
Registered: February 2000
|
Member |
|
|
Hmmm.... Well assuming that you want to iterate (for whatever reason)
over a set of names rather than a set of indices how about.
names = tagnames(data)
for i=0,n_tags(data)-1 do begin ; Or any other kind of loop
name = ... entered somehow by the user perhaps ...
w=where(name eq names[i]);
if (w(0) ne -1) then begin
ind = w[0]
data.(ind).values = ...
....
It's easy enough to translate from the variable name to the variable
index -- and the index may itself be a variable.
Tom
Randall Skelton wrote:
>
> Thanks Tom... I did think of that. However, in this particular case there
> is some merit in having the name of the structure be a useful and human
> readable tag. Nobody would be happy trying to remember yet another
> arbitrary numbering scheme for molecules when they'd rather just type the
> name ;) My suspicion is that there isn't an easy way to do what I want...
>
> Randall
>
> On Thu, 19 Apr 2001, tam wrote:
>
>> Just use the numeric tag syntax for structures.
>>
>> for i=0,n_tags(data)-1 do begin
>> data.(i).values = .. some expresion ...
>> endfor
>>
>> where the (i) indicates use the i'th element of the structure.
>> So you don't need to use the names at all.
>>
>> Regards,
>> Tom McGlynn
>>
|
|
|
|
|
|
| Re: indexing over structure tags [message #24807 is a reply to message #24802] |
Thu, 19 April 2001 11:08  |
Paul van Delst
Messages: 364
Registered: March 1997
|
Senior Member |
|
|
Randall Skelton wrote:
>
> Thanks Tom... I did think of that. However, in this particular case there
> is some merit in having the name of the structure be a useful and human
> readable tag. Nobody would be happy trying to remember yet another
> arbitrary numbering scheme for molecules when they'd rather just type the
> name ;) My suspicion is that there isn't an easy way to do what I want...
Maybe more context is needed to solve your problem .... the user of the code *shouldn't*
have to remember the (not so) arbitrary numbering scheme - the user would type in a
molecule name (or names). How your code deals with searching the human readable tagnames
is a different matter, no?
e.g.
; define the basic structure for each
sm_struc = {basic_struct, comment: ' ', values: fltarr(nlev)}
; define the large structure
data = {big_struct, so4: sm_struc, co2: sm_struc, hcl: sm_struc}
; get the names of the tags
names = tag_names(data)
so that names = [so4, co2, hcl].
Say the user requests data for 'so4' and 'hcl' so how about
user_request = ['so4', 'hcl']
n_requests = N_ELEMENTS( user_request )
FOR i = 0, n_requests - 1 DO BEGIN
tag_number = (WHERE( user_request[i] EQ names ))[0] ; <-- assume this always succeeds
data_to_get = data.(tag_number).values
IF ( i EQ 0 ) THEN $
data_to_return = data_to_get $
ELSE $
data_to_return = [ [ data_to_return ], [ data_to_get ] ]
ENDFOR
RETURN, data_to_return
or something like that? As it is above might not work for plucking out structures, but
that's a detail. So is the concatenation build of the data_to_return. Should be o.k. for
small arrays tho'.
paulv
--
Paul van Delst A little learning is a dangerous thing;
CIMSS @ NOAA/NCEP Drink deep, or taste not the Pierian spring;
Ph: (301)763-8000 x7274 There shallow draughts intoxicate the brain,
Fax:(301)763-8545 And drinking largely sobers us again.
paul.vandelst@noaa.gov Alexander Pope.
|
|
|
|
| Re: indexing over structure tags [message #24809 is a reply to message #24807] |
Thu, 19 April 2001 10:36 |
Vapuser
Messages: 63
Registered: November 1998
|
Member |
|
|
Randall Skelton <rhskelto@atm.ox.ac.uk> writes:
> Hello,
>
> Imagine someone has a structure of structures...
>
> ; define the basic structure for each
> sm_struc = {basic_struct, comment: ' ', values: fltarr(nlev)}
>
> ; define the large structure
> data = {big_struct, so4: sm_struc, co2: sm_struc, hcl: sm_struc}
>
> The IDL manual describes how to make an array of the tags in a structure
> using:
>
> ; get the names of the tags
> names = tag_names(data)
>
> so that names = [so4, co2, hcl].
>
> That is all fine. But is it possible to index over the tag names with a
> for-loop?
>
> i.e. for i=0, n_elements(names)-1 do data.names[i].values = i
> ^^^^^^^^^^^^^^^^^^^^
>
> where IDL determines what the appropriate label 'data.name[i].values' is.
>
> Thanks,
> Randall
>
data.(i) to iterate over that tags of data, and
data.(i).(j) to iterate over the tags of data.(i)
If each tag is an array, rather than a scalar structure instance, you
might (probably?) have to pull it out, like this
IDL> junk={foo, a:0, b:0l, c:0.0}
IDL> bar={bar, a:replicate({foo},3),b:replicate({foo},4)}
IDL> help,bar,/struct
** Structure BAR, 2 tags, length=84:
A STRUCT -> FOO Array[3]
B STRUCT -> FOO Array[4]
IDL> .run
- FOR i=0,1 DO BEGIN
- FOR j=0,2 DO BEGIN
- tmp=bar.(i).(j) ;<--- I couldn't find a way to do this inline,
; ; but there might be one that I'm missing
- FOR k=0,n_elements(tmp)-1 DO tmp[k]=(2^(i+1))*(3^(j+1))+k
- bar.(i).(j)=tmp
- endfor
- endfor
- end
% Compiled module: $MAIN$.
IDL> print,bar
{
a = { 6 18 54.0000}
{ 7 19 55.0000}
{ 8 20 56.0000}
b = { 12 36 108.000}
{ 13 37 109.000}
{ 14 38 110.000}
{ 15 39 111.000}
}
whd
--
William Daffer: 818-354-0161: William.Daffer@jpl.nasa.gov
|
|
|
|
| Re: indexing over structure tags [message #24810 is a reply to message #24809] |
Thu, 19 April 2001 10:50 |
Bernard Puc
Messages: 65
Registered: January 1998
|
Member |
|
|
Randall Skelton wrote:
>
> Thanks Tom... I did think of that. However, in this particular case there
> is some merit in having the name of the structure be a useful and human
> readable tag. Nobody would be happy trying to remember yet another
> arbitrary numbering scheme for molecules when they'd rather just type the
> name ;) My suspicion is that there isn't an easy way to do what I want...
>
> Randall
>
> On Thu, 19 Apr 2001, tam wrote:
>
>> Just use the numeric tag syntax for structures.
>>
>> for i=0,n_tags(data)-1 do begin
>> data.(i).values = .. some expresion ...
>> endfor
>>
>> where the (i) indicates use the i'th element of the structure.
>> So you don't need to use the names at all.
>>
>> Regards,
>> Tom McGlynn
>>
Perhaps you misunderstood...there are two ways of accessing a tag within
a structure.
data = {lat:fltarr(10), lng:fltarr(10), sig:fltarr(10)}
Method 1:
data.lat
data.lng
data.sig
Method 2:
data.(0)
data.(1)
data.(2)
Either method accesses the same variables.
--
Bernard Puc AETC, INC.
bpuc@va.aetc.com 1225 Jefferson Davis Highway #800
(703) 413-0500 Arlington, VA 22202
|
|
|
|
| Re: indexing over structure tags [message #24811 is a reply to message #24809] |
Thu, 19 April 2001 10:16 |
tam
Messages: 48
Registered: February 2000
|
Member |
|
|
Just use the numeric tag syntax for structures.
for i=0,n_tags(data)-1 do begin
data.(i).values = .. some expresion ...
endfor
where the (i) indicates use the i'th element of the structure.
So you don't need to use the names at all.
Regards,
Tom McGlynn
Randall Skelton wrote:
>
> Hello,
>
> Imagine someone has a structure of structures...
>
> ; define the basic structure for each
> sm_struc = {basic_struct, comment: ' ', values: fltarr(nlev)}
>
> ; define the large structure
> data = {big_struct, so4: sm_struc, co2: sm_struc, hcl: sm_struc}
>
> The IDL manual describes how to make an array of the tags in a structure
> using:
>
> ; get the names of the tags
> names = tag_names(data)
>
> so that names = [so4, co2, hcl].
>
> That is all fine. But is it possible to index over the tag names with a
> for-loop?
>
> i.e. for i=0, n_elements(names)-1 do data.names[i].values = i
> ^^^^^^^^^^^^^^^^^^^^
>
> where IDL determines what the appropriate label 'data.name[i].values' is.
>
> Thanks,
> Randall
|
|
|
|
|
|
| Re: indexing over structure tags [message #24813 is a reply to message #24811] |
Thu, 19 April 2001 10:05 |
Paul van Delst
Messages: 364
Registered: March 1997
|
Senior Member |
|
|
Paul van Delst wrote:
>
> Randall Skelton wrote:
>>
>> Hello,
>>
>> Imagine someone has a structure of structures...
>>
>> ; define the basic structure for each
>> sm_struc = {basic_struct, comment: ' ', values: fltarr(nlev)}
>>
>> ; define the large structure
>> data = {big_struct, so4: sm_struc, co2: sm_struc, hcl: sm_struc}
>>
>> The IDL manual describes how to make an array of the tags in a structure
>> using:
>>
>> ; get the names of the tags
>> names = tag_names(data)
>>
>> so that names = [so4, co2, hcl].
>>
>> That is all fine. But is it possible to index over the tag names with a
>> for-loop?
>>
>> i.e. for i=0, n_elements(names)-1 do data.names[i].values = i
>> ^^^^^^^^^^^^^^^^^^^^
>
> Maybe
>
> data.names[i].(j) where j=0->N_TAGS(data.names[i])-1 ??
>
> This works on "single-level" structures but I don't see why it wouldn't work on nested
> ones.
Oops, I guess you really wanted something like
data.(i).values
where i=0->number of molecular names
'Scusi
paulv
--
Paul van Delst A little learning is a dangerous thing;
CIMSS @ NOAA/NCEP Drink deep, or taste not the Pierian spring;
Ph: (301)763-8000 x7274 There shallow draughts intoxicate the brain,
Fax:(301)763-8545 And drinking largely sobers us again.
paul.vandelst@noaa.gov Alexander Pope.
|
|
|
|
| Re: indexing over structure tags [message #24814 is a reply to message #24813] |
Thu, 19 April 2001 10:00 |
Paul van Delst
Messages: 364
Registered: March 1997
|
Senior Member |
|
|
Randall Skelton wrote:
>
> Hello,
>
> Imagine someone has a structure of structures...
>
> ; define the basic structure for each
> sm_struc = {basic_struct, comment: ' ', values: fltarr(nlev)}
>
> ; define the large structure
> data = {big_struct, so4: sm_struc, co2: sm_struc, hcl: sm_struc}
>
> The IDL manual describes how to make an array of the tags in a structure
> using:
>
> ; get the names of the tags
> names = tag_names(data)
>
> so that names = [so4, co2, hcl].
>
> That is all fine. But is it possible to index over the tag names with a
> for-loop?
>
> i.e. for i=0, n_elements(names)-1 do data.names[i].values = i
> ^^^^^^^^^^^^^^^^^^^^
Maybe
data.names[i].(j) where j=0->N_TAGS(data.names[i])-1 ??
This works on "single-level" structures but I don't see why it wouldn't work on nested
ones.
paulv
--
Paul van Delst A little learning is a dangerous thing;
CIMSS @ NOAA/NCEP Drink deep, or taste not the Pierian spring;
Ph: (301)763-8000 x7274 There shallow draughts intoxicate the brain,
Fax:(301)763-8545 And drinking largely sobers us again.
paul.vandelst@noaa.gov Alexander Pope.
|
|
|
|
| Re: indexing over structure tags [message #24871 is a reply to message #24812] |
Tue, 24 April 2001 00:45 |
R.Bauer
Messages: 1424
Registered: November 1998
|
Senior Member |
|
|
Randall Skelton wrote:
>
> Thanks Tom... I did think of that. However, in this particular case there
> is some merit in having the name of the structure be a useful and human
> readable tag. Nobody would be happy trying to remember yet another
> arbitrary numbering scheme for molecules when they'd rather just type the
> name ;) My suspicion is that there isn't an easy way to do what I want...
>
> Randall
>
Dear Randall,
I will show you an easy way, how to access any level of a structure
by incrementing of a an array.
may be we have a structure like
pi={name:'p.mustermann'}
global={pi:pi}
n2o={param:fltarr(10),short_name:'N2O',long_name:'mixing
ratio',units:'ppb',flag:'NONE'}
time={param:fltarr(10),short_name:'time',long_name:'time
[UT]',units:'seconds since 2000-01-01 00:00:00 UTC', flag:'NONE'}
icgs={!global:global,time:time,n2o:n2o}
more attribute of this data structure are explained by my publication.
http://www.fz-juelich.de/zb/text/publikation/juel3786.html
The following routine creates an array of tagnames and an
array of pointers of any submitted structure
The return value of this function is n_elements(names)
and the keywords names and ptr_values
n=struct2names_and_ptrs(icgs,names=names,ptr_values=ptr_valu es)
PRINT,n
IDL> 11
IDL> PRINT,names
!GLOBAL.PI.NAME TIME.PARAM TIME.SHORT_NAME TIME.LONG_NAME TIME.UNITS
TIME.FLAG
N2O.PARAM N2O.SHORT_NAME N2O.LONG_NAME N2O.UNITS N2O.FLAG
IDL> FOR I=0,n-1 do help,*(ptr_values[i])
<PtrHeapVar1> STRING = 'p.mustermann'
<PtrHeapVar2> FLOAT = Array[10]
<PtrHeapVar3> STRING = 'time'
<PtrHeapVar4> STRING = 'time [UT]'
<PtrHeapVar5> STRING = 'seconds since 2000-01-01 00:00:00 UTC'
<PtrHeapVar6> STRING = 'NONE'
<PtrHeapVar7> FLOAT = Array[10]
<PtrHeapVar8> STRING = 'N2O'
<PtrHeapVar9> STRING = 'mixing ratio'
<PtrHeapVar10> STRING = 'ppb'
<PtrHeapVar11> STRING = 'NONE'
http://www.fz-juelich.de/icg/icg1/idl_icglib/idl_source/idl_ html/dbase/download/struct2names_and_ptrs.tar.gz
This routine was written to use by add_tag, delete_tag, replace_tagvalue
and rename_tag
on each level of a structure with the following routine you are able to
build any structure defined from names and ptr_values.
http://www.fz-juelich.de/icg/icg1/idl_icglib/idl_source/idl_ html/dbase/download/names_and_ptrs2struct.tar.gz
For further routines and licensing please look at
http://www.fz-juelich.de/icg/icg1/idl_icglib/idl_lib_intro.h tml
Is this the solution of your problem?
Reimar
--
Reimar Bauer
Institut fuer Stratosphaerische Chemie (ICG-1)
Forschungszentrum Juelich
email: R.Bauer@fz-juelich.de
http://www.fz-juelich.de/icg/icg1/
=============================================
a IDL library at ForschungsZentrum J�lich
http://www.fz-juelich.de/icg/icg1/idl_icglib/idl_lib_intro.h tml
http://www.fz-juelich.de/zb/text/publikation/juel3786.html
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
