writing sequences to a file
From: Bart Vandewoestyne (MyFirstName.MyLastName_at_telenet.be)
Date: 03/29/05
- Previous message: glen herrmannsfeldt: "Re: About portable programming in fortran"
- Next in thread: glen herrmannsfeldt: "Re: writing sequences to a file"
- Reply: glen herrmannsfeldt: "Re: writing sequences to a file"
- Reply: Richard E Maine: "Re: writing sequences to a file"
- Reply: Bill C: "Re: writing sequences to a file"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Tue, 29 Mar 2005 08:31:10 +0000 (UTC)
Hello,
Suppose that in your work, you are working with sequences of s-dimensional
points. The sequences have N such points. Both s and N can become very
large (say about up to 1000 or 10000 for s, and up to 10 million for N).
All numbers are real values between 0 and 1. I want to write these
points to a file using as much precision as possible.
The parameters s and N are only known at run-time, so most of the time
you work with allocatable arrays for which the size depends on s and N.
What would be the best format to write such a N x s sequence to a file
so you can easily compare it with a sequence somebody else sends to you
in the same format (checking for equality, e.g. the numbers could be N
s-dimensional points generated by a random number generator).
The main 'problem' is that i only know s at runtime, so I don't know
what number to use as my repeat count in my edit descriptors.
It would be nice if i could do something like this for an array 'sequence'
consisting of 'row_number' rows and 's' columns:
write(unit=unit_number, fmt="(sES20.15)") sequence(row_number, :)
but this has two problems:
1) I can't use the variable 's' as a repeat count for my edit descriptor
2) Is there a bound on the repeat count or can i use as many repeat
counts as I want? Because if s is 1000 or 1000, that would mean i am
using *a lot* of columns in my output file... i remember having tried
this once, and then i got some error that i used to many columns or
something like that... don't know the exact details anymore though... is
there actually a bound on the number of columns in an output file or
should it be possible to write say about 10000 reals with 15 decimals in
one row/line of your output file?
Another option would be to use a maximum of say 10 columns and spread
one s-dimensional point over multiple lines, and separate the points
with an empty line. For example for 15 dimensional sequences:
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
0.11 0.12 0.13 0.14 0.15
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
0.11 0.12 0.13 0.14 0.15
... and so on...
How would you people do this? What would be the best option to store
the points so you can also easily compare them with the points from
somebody else who sticks to the same output format?
Regards,
Bart
-- "Share what you know. Learn what you don't."
- Previous message: glen herrmannsfeldt: "Re: About portable programming in fortran"
- Next in thread: glen herrmannsfeldt: "Re: writing sequences to a file"
- Reply: glen herrmannsfeldt: "Re: writing sequences to a file"
- Reply: Richard E Maine: "Re: writing sequences to a file"
- Reply: Bill C: "Re: writing sequences to a file"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
|
