| Splitting An Array Of Strings Without Using Loops [message #35881] |
Thu, 24 July 2003 21:42  |
darrick.white
Messages: 7
Registered: January 2003
|
Junior Member |
|
|
This is probably simple, but I'm having a time trying to figure it
out. I want to be able to split an array of strings without using
loops.
Example:
dataPoints is an array of strings with N elements
The format of each element within dataPoints is "x:y1:y2:y3:yn". More
than likely, the data will be in the format of x:y".
This array will become data points (the first element is always
considered the x coordinate): (x,y) = 1,23. In case of multiple
points (2:21:34:54), the data will look like: (2,21), (2,34), (2,54).
I need a way to take:
dataPoints[0] = 1:23
dataPoints[1] = 2:32
dataPoints[2] = 3:30
dataPoints[3] = 4:45
and create
points[2,4]
1 23
2 32
3 30
4 45
-Darrick
|
|
|
|
|
|
| Re: Splitting An Array Of Strings Without Using Loops [message #35929 is a reply to message #35881] |
Tue, 29 July 2003 09:24  |
JD Smith
Messages: 850
Registered: December 1999
|
Senior Member |
|
|
On Tue, 29 Jul 2003 08:46:00 -0700, Pavel Romashkin wrote:
> JD Smith wrote:
>>
>> Come on people. I don't use HISTOGRAM for everything. I use it very
>> rarely, in fact.
>
> We don't belive this! Now that you've got the reputation, there is no
> getting away from it :-)
>
>> You could (yes) use HISTOGRAM or perhaps many other methods
>
> See? Told you! :-)
*You* could use HISTOGRAM. I've given it up for the month.
JD
|
|
|
|
| Re: Splitting An Array Of Strings Without Using Loops [message #35931 is a reply to message #35881] |
Tue, 29 July 2003 08:46 |
Pavel Romashkin
Messages: 166
Registered: April 1999
|
Senior Member |
|
|
JD Smith wrote:
>
> Come on people. I don't use HISTOGRAM for everything. I use it very
> rarely, in fact.
We don't belive this! Now that you've got the reputation, there is no
getting away from it :-)
> You could (yes) use HISTOGRAM or perhaps many other methods
See? Told you! :-)
Cheers,
Pavel
|
|
|
|
| Re: Splitting An Array Of Strings Without Using Loops [message #35936 is a reply to message #35881] |
Mon, 28 July 2003 16:00 |
JD Smith
Messages: 850
Registered: December 1999
|
Senior Member |
|
|
On Mon, 28 Jul 2003 11:24:50 -0700, Rick Towler wrote:
> "Darrick White" wrote...
>
>> It looks like I'm not explaining my problem clearly.
>
>> Is there a way (not knowing what data set input is used) to transform
>> my data into the corresponding result array?
>
> I don't think the issue is one of clarity, but of possibility. Unless
> JD can save you with some magical incarnation of HISTOGRAM you are going
> to have to change your design criteria or use a loop. If performance is
> really that important write this function in C.
>
> -Rick
Come on people. I don't use HISTOGRAM for everything. I use it very
rarely, in fact.
How about something like:
nums=strsplit(strjoin(data,':'),':',/EXTRACT)
cnts=long(total(byte(data) eq 58b,1))+1L
Now you have a list of tuple-counts and the tuples themselves in a long
list. You could (yes) use HISTOGRAM or perhaps many other methods to
stick these into an array as you describe without looping, but rather than
show something you'd forget 5 minutes after dropping it into your code,
I'll join Rick in saying that if parsing these strings quickly is this
important to you, you'll get better results by re-designing the input
format, or pre-parsing them using a language better suited to these
manipulations. And on the off chance that you're suffering from the
"must-optimize-everything-in-sight" disease, you'll want to make sure a
readable and straightforward input loop won't meet your needs before
venturing too far into IDL esoterica:
b=make_array(/LONG,VALUE=-1,max(cnt),n_elements(data))
for i=0,n_elements(data)-1 do b[0,i]=strsplit(data[i],':',/EXTRACT)
Note that there's no integer (long or otherwise) definition of NaN, so I
used -1.
JD
|
|
|
|
| Re: Splitting An Array Of Strings Without Using Loops [message #35942 is a reply to message #35881] |
Mon, 28 July 2003 11:47 |
Paul Van Delst[1]
Messages: 1157
Registered: April 2002
|
Senior Member |
|
|
Darrick White wrote:
>
>> Dear Darrick,
>>
>> here is a second solution using reads.
>>
>> pro test
>> data=['1:23','2:32','3:30','4:45']
>>
>> s={x:bytarr(1),s:bytarr(1),y:bytarr(2)}
>> s=replicate(s,4)
>>
>> reads,byte(data),s
>>
>> print,string(s.x)
>> print,string(s.y)
>> end
>>
>> IDL> 1 2 3 4
>> IDL> 23 32 30 45
>
> It looks like I'm not explaining my problem clearly. For instance,
> the following sets of data are valid inputs to my application:
>
> 1) data=['1:23','2:32','3:30','4:45']
> 2) data=['12:23','22:32:34:45','32:30','42:45:90']
> 3) data=['100:23','200:32','300:30','400:45']
> 4) data=['1:23:2','2:32:2','3:30:2','4:45:2']
>
> The resulting transformation would like this for both:
>
> 1) print, intarr(2,4)
> 1 23
> 2 32
> 3 30
> 4 45
>
> 2) print, intarr(4,4)
> 12 23 NaN NaN
> 22 32 34 45
> 32 30 NaN NaN
> 42 45 90 NaN
Wouldn't this need to be a two-pass problem? You parse the input data to
determine the individual entry and maximum dimension (in this case 4 due to the
22:32:34:45), create you array with fill values, and then "go through the array
once more" to fill in your array. (The quotes are there because going through
the array once more could be achieved a number of ways.)
I would think that smart usage of the IDL string functions should be able to do
most of that sans looping. (Otherwise, I'm sure JD can come up with some neato
supa-quick method using HISTOGRAM.... :o)
paulv
p.s. If you're only using integers, you can't use NaN as a fill value.
--
Paul van Delst
CIMSS @ NOAA/NCEP/EMC
Ph: (301)763-8000 x7748
Fax:(301)763-8545
|
|
|
|
| Re: Splitting An Array Of Strings Without Using Loops [message #35943 is a reply to message #35881] |
Mon, 28 July 2003 11:24 |
Rick Towler
Messages: 821
Registered: August 1998
|
Senior Member |
|
|
"Darrick White" wrote...
> It looks like I'm not explaining my problem clearly.
> Is there a way (not knowing what data set input is used) to transform
> my data into the corresponding result array?
I don't think the issue is one of clarity, but of possibility. Unless JD
can save you with some magical incarnation of HISTOGRAM you are going to
have to change your design criteria or use a loop. If performance is really
that important write this function in C.
-Rick
|
|
|
|
|
|
| Re: Splitting An Array Of Strings Without Using Loops [message #35946 is a reply to message #35881] |
Mon, 28 July 2003 10:17 |
darrick.white
Messages: 7
Registered: January 2003
|
Junior Member |
|
|
> Dear Darrick,
>
> here is a second solution using reads.
>
> pro test
> data=['1:23','2:32','3:30','4:45']
>
> s={x:bytarr(1),s:bytarr(1),y:bytarr(2)}
> s=replicate(s,4)
>
> reads,byte(data),s
>
> print,string(s.x)
> print,string(s.y)
> end
>
> IDL> 1 2 3 4
> IDL> 23 32 30 45
It looks like I'm not explaining my problem clearly. For instance,
the following sets of data are valid inputs to my application:
1) data=['1:23','2:32','3:30','4:45']
2) data=['12:23','22:32:34:45','32:30','42:45:90']
3) data=['100:23','200:32','300:30','400:45']
4) data=['1:23:2','2:32:2','3:30:2','4:45:2']
The resulting transformation would like this for both:
1) print, intarr(2,4)
1 23
2 32
3 30
4 45
2) print, intarr(4,4)
12 23 NaN NaN
22 32 34 45
32 30 NaN NaN
42 45 90 NaN
3) print, intarr(2,4)
100 23
200 32
300 30
400 45
4) print, intarr(3,4)
1 23 2
2 32 2
3 30 2
4 45 2
Is there a way (not knowing what data set input is used) to transform
my data into the corresponding result array? Note: For
transformation #2 above, I need to append each point to my new array.
If the array dimensions don't match, I need to fill in those missing
elements with 'NaN'.
Thanks
-Darrick
|
|
|
|
| Re: Splitting An Array Of Strings Without Using Loops [message #35960 is a reply to message #35881] |
Sat, 26 July 2003 04:10 |
R.Bauer
Messages: 1424
Registered: November 1998
|
Senior Member |
|
|
Darrick White wrote:
> This is probably simple, but I'm having a time trying to figure it
> out. I want to be able to split an array of strings without using
> loops.
>
> Example:
> dataPoints is an array of strings with N elements
> The format of each element within dataPoints is "x:y1:y2:y3:yn". More
> than likely, the data will be in the format of x:y".
>
> This array will become data points (the first element is always
> considered the x coordinate): (x,y) = 1,23. In case of multiple
> points (2:21:34:54), the data will look like: (2,21), (2,34), (2,54).
>
> I need a way to take:
> dataPoints[0] = 1:23
> dataPoints[1] = 2:32
> dataPoints[2] = 3:30
> dataPoints[3] = 4:45
>
>
> and create
> points[2,4]
> 1 23
> 2 32
> 3 30
> 4 45
>
> -Darrick
Dear Darrick,
here is a second solution using reads.
pro test
data=['1:23','2:32','3:30','4:45']
s={x:bytarr(1),s:bytarr(1),y:bytarr(2)}
s=replicate(s,4)
reads,byte(data),s
print,string(s.x)
print,string(s.y)
end
IDL> 1 2 3 4
IDL> 23 32 30 45
--
Forschungszentrum Juelich
email: R.Bauer@fz-juelich.de
http://www.fz-juelich.de/icg/icg-i/
============================================================ ======
a IDL library at ForschungsZentrum Juelich
http://www.fz-juelich.de/icg/icg-i/idl_icglib/idl_lib_intro. html
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
