I would like to write data into hdf5 file in fortran in an appended way, but I can't hold the data in a large memory array.
when writing using ascii I can do something like:
program example1
implicit none
integer :: nu, i
open(filename="someFile.txt", newunit=nu)
do i = 1, 10
write(nu, *) i
end do
close(nu)
end program example1
and what I would get is a file holding all of the integer numbers from 1 to 10 but without declaring a size 10 one dimensional array.
How can I write a hdf5 file in fortran that does exactly that, i.e holding the numbers in the same file under the same dataset name but without holding an array that holds these numbers?
You need to create an extendible dataset and then populate it with the help of an hyperslab or point selection (otherwise, previously written data will be overwritten). Using HDFql, your use-case can be solved as follows in Fortran:
PROGRAM Example
USE HDFql
CHARACTER :: variable_number
INTEGER :: state
INTEGER :: i
state = hdfql_execute("CREATE AND USE FILE test.h5")
state = hdfql_execute("CREATE DATASET dset AS INT(0 TO UNLIMITED)")
WRITE(variable_number, "(I0)") hdfql_variable_register(i)
DO i = 1, 10
state = hdfql_execute("ALTER DIMENSION dset TO +1")
state = hdfql_execute("INSERT INTO dset(-1) VALUES FROM MEMORY " // variable_number)
END DO
state = hdfql_variable_unregister(i)
state = hdfql_execute("CLOSE FILE")
END PROGRAM
Related
I am trying to read a text file using a Fortran code. I have a file with 1999 rows and the number of columns vary with each row. Can someone please tell me how one can code such a problem. This is my code for reading a 4*2 text file but I am using do loops which I can't use in my current case.
PROGRAM myread2
IMPLICIT NONE
INTEGER, DIMENSION(100) :: a, b
INTEGER :: row,col,max_rows,max_cols
OPEN(UNIT=11, file='text.txt')
DO row = 1,4
READ(11,*) a(row), b(row)
END DO
PRINT *, a(1)
PRINT *, a(4)
PRINT*, b(4)
END PROGRAM myread2
The best way of reading a file like this depends on how you want to store the data. I'm going to use a ragged array as it's probably simplest, although other container types may be better suited depending on your requirements.
Fortran doesn't have ragged arrays natively, so first you need to define a type to hold each row. This can be done as
type :: RowData
integer, allocatable :: cols(:)
end type
type(RowData), allocatable :: rows(:)
When this container is filled out, the value in the i'th column of the j'th row will be accessed as
value = rows(j)%cols(i)
We can then write a program to read the file, e.g.
type :: RowData
integer, allocatable :: cols(:)
end type
type(RowData), allocatable :: rows(:)
integer :: no_rows
integer :: i
open(unit=11, file='text.txt')
no_rows = count_lines(11)
allocate(rows(no_rows))
do i=1,no_rows
rows(i)%cols = read_row(11)
enddo
Now we just need to write the functions count_lines, which counts the number of lines in the file, and read_row, which reads a line from the file and returns the contents of that line as an array of integers.
Following this question, count_lines can be written as
! Takes a file unit, and returns the number of lines in the file.
! N.B. the file must be at the start of the file.
function count_lines(file_unit) result(output)
integer, intent(in) :: file_unit
integer :: output
integer :: iostat
output = 0
iostat = 0
do while (iostat==0)
read(file_unit, *, iostat=iostat)
if (iostat==0) then
output = output+1
endif
enddo
rewind(file_unit)
end function
Writing read_row, to parse a line of unknown length from a file, can be done by following this question.
This question already has answers here:
Passing character strings of different lengths to functions in Fortran
(2 answers)
Closed 2 years ago.
I have a bunch of files, which contain numbers in rows. I need to write a function, that
reads from file for the first time to find amount of elements in file;
allocates an array and reads numbers from file into array;
returns an array
My function gets a string - name of the file - as an input.
So, the function that I wrote is:
function arrays_proc(name) result(arr)
character(len=128), intent(in) :: name
integer :: i, tmp, ios
character(len=30) :: line
double precision, dimension(:), allocatable :: arr
open(unit=09, file=name, status='old', iostat=ios)
if ( ios /= 0) stop "error opening file"
tmp = 0
do
read(09, '(A)', iostat=ios) line
if (ios /= 0) exit
tmp = tmp + 1
end do
allocate(arr(tmp))
rewind(09)
do i=1, tmp
read(09, '(A)') arr(i)
end do
close(09)
return
end function arrays_proc
Then, in the main program I write
...
real(8), dimension(:), allocatable :: points, potent
points = arrays_proc(trim('carbon_mesh.txt'))
potent = arrays_proc(trim('carbon_pot.txt'))
...
When I run my program, I get instant "error opening file".
I assume the problem is with names of files or how I put them in my function.
Anyway, I hope someone can help me
When compiling your code with a minimal program, GFortran prints the following warnings:
a.f90:4:25:
4 | points = arrays_proc(trim('carbon_mesh.txt'))
| 1
Warning: Character length of actual argument shorter than of dummy argument ‘name’ (15/128) at (1)
a.f90:5:25:
5 | potent = arrays_proc(trim('carbon_pot.txt'))
| 1
Warning: Character length of actual argument shorter than of dummy argument ‘name’ (14/128) at (1)
Trying to print the value of name inside arrays_proc shows that it is filled with garbage. So, guided by the warnings, you can try to change the length of the name parameter to *, which allows a string of any length to be used as input.
With that change, the function manages to open the files.
See also: Passing character strings of different lengths to functions in Fortran
I am working on a flood-fill recursion assignment where I have to read an ASCII art text file and fill it in. The assignment can be found here: https://faculty.utrgv.edu/robert.schweller/CS2380/homework/hw10.pdf
Recursion() //construcor
{
column = -1;
row = -1;
grid = new char*[size of art row];
for(int i = 0; i < size of art; i++)
{
board[row] = new char[size of art column]
}
}
I'm not sure if determining the size of the array should be in the constructor or not. I need to know the size of the array in order to know where the user wants to fill the art file in. Also, here is all of the code for a better context. https://pastebin.com/TSYH26Ci
I would handle your file as binary. I do not know which OS,API you are using so I will answer just in generic way...
get file size siz
usually seeking to 0 bytes from end of file will give you the file size.
allocate 1D array dat to store your entire file
dat = new BYTE[siz];
load your file into memory (1D array)
do not forget to use binary access as some ASCII arts could use control codes (ASCII below 32) which could be corrupted by text file access.
scan for end of line
so scan your array from 0 and stop when you find ASCII codes 13 or 10. Its position will give you the x resolution of your ASCII art
int xs;
for (xs=0;xs<siz;xs++)
if ((dat[xs]==10)||(dat[xs]==13))
break;
now xs should be holding your x resolution.
compute y resolution ys
the safest way is to count the number of end of lines (13 or 10). In such case you can even store the line start addresses to some pointer array BYTE **pixel=new (BYTE*)[ys]; which will enable you simple 2D access pixel[y][x]. If your ASCII art is aligned and have constant size per each line than you can compute ys from size..
ys = siz/(xs+eol_size)
where eol_size is 1 or 2 depending on the used line ending: ((10),(13),(13,10) or (10,13)) so:
eol_size=1;
if (xs<siz)
if ((dat[xs+1]==10)||(dat[xs+1]==13))
eol_size=2;
As we do not have access to any input file we can only guess... If you need to generate one see:
C++ Image to ASCII art conversion
Here example of binary file access in VCL (bullets #1,#2,#3):
Convert the Linux open, read, write, close functions to work on Windows
I am having trouble reading exponential from a text file using Fortran.
The entry in the text file looks like the following
0.02547163e+06-0.04601176e+01 0.02500000e+02 0.00000000e+00 0.00000000e+00 3
And the code that I am using looks like the following
read(iunit,'(ES20.8,ES20.8,ES20.8,ES20.8,ES20.8,I2)') dummy1, dummy2, Thermo_DB_Coeffs_LowT(iS,1:3),temp
The error I am getting is
Fortran runtime error: Bad value during floating point read
How can I read these values?
Well here is what I usually do when it is too painful to hand edit the file...
CHARACTER(LEN=256) :: Line
INTEGER, PARAMETER :: Start = 1
INTEGER :: Fin, Trailing_Int, I
DOUBLE, DIMENSION(6) :: Element
...
Ingest_All_Rows: DO WHILE(.TRUE.)
READ(...) Line ! Into a character-string
<if we get to the end of the file, then 'EXIT Ingest_All_Rows'>
Start =1
Single_Row_Ingest: DO I = 1, 6
Fin = SCAN(Line,'eE')+3 !or maybe it is 4?
IF(I ==6) Fin = LEN_TRIM(Line)
READ(Line(Start:Fin),*) Element(I) !fron the string(len-string) to the double.
Line = Line((Fin+1):)
IF(I ==6) Trailing_Int = Element(6)
ENDDO Single_Row_Ingest
<Here we shove the row's 5 elements into some array, and the trailing int somewhere>
ENDDO Ingest_All_Rows
You will have to fill in the blanks, but I find that SCAN and LEN_TRIM can be useful in these cases
I am currently writing a code to simulate particle collisions. I am trying to open as much files as there are particles (N) and then put the data for positions and velocities in each of these files for each step of the time integration (using Euler's method, but that is not relevant). For that, I tried using a do loop so it will open all the files I need - then I put all the data in them with a different do loop later - and then close them all.
I first tried just doing a do loop to open the files - but it gave errors of the type "file already open in another unit", so I did the following:
module parameters
implicit none
character :: posvel
integer :: i, j, N
real :: tmax
real, parameter :: tmin=0.0, pi=3.14159265, k=500.0*10E3, dt=10.0E-5, dx=10.0E-3, g=9.806, ro=1.5*10E3
real, dimension(:), allocatable :: xold, xnew, vold, vnew, m, F, r
end module parameters
PROGRAM Collision
use parameters
implicit none
write(*,*) 'Enter total number of particles (integer number):'
read(*,*) N
allocate(xold(N))
allocate(vold(N))
allocate(xnew(N))
allocate(vnew(N))
allocate(m(N))
allocate(F(N))
allocate(r(N))
xold(1) = 0.0
vold(1) = 0.0
m(1) = 6.283*10E-9
r(1) = 10E-4
xold(2) = 5.0
vold(2) = 0.0
m(2) = 6.283*10E-9
r(2) = 10E-4
write(*,*) 'Type total time elapsed for the simulation(real number):'
read(*,*) tmax
do i = 1, N
write(posvel,"(a,i3.3,a)") "posveldata",i,".txt"
open(unit=i,file=posvel, status="unknown")
end do
do i = 1, N
close(unit=i)
end do
END PROGRAM Collision
The last ten lines are the ones that regard to my problem.
That worked in codeblocks - it opened just the number of files I needed, but I'm actually using gfortran and it gives me and "end of record" error in the write statement.
How can I make it to execute properly and give me the N different files that I need?
P.S.: It is not a problem of compilation, but after I execute the program.
Your character string in the parameter module has only 1 character length, so it cannot contain the full file name. So please use a longer string, for example
character(100) :: posvel
Then you can open each file as
do i = 1, N
write(posvel,"(a,i0,a)") "posveldata",i,".txt"
open(unit=i,file=trim(posvel), status="unknown")
end do
Here, I have used the format i0 to automatically determine a proper width for integer, and trim() for removing unnecessary blanks in the file name (though they may not be necessary). The write statement can also be written more compactly as
write(posvel,"('posveldata',i0,'.txt')") i
by embedding character literals into the format specification.
The error message "End of record" comes from the above issue. This can be confirmed by the following code
character c
write(c,"(a)") "1"
print *, "c = ", c
write(c,"(a)") "23" !! line 8 in test.f90
print *, "c = ", c
for which gfortran gives
c = 1
At line 8 of file test.f90
Fortran runtime error: End of record
This means that while c is used as an internal file, this "file" does not have enough space to accommodate two characters (here "23"). For comparison, ifort14 gives
c = 1
forrtl: severe (66): output statement overflows record, unit -5, file Internal Formatted Write
while Oracle Fortran12 gives
c = 1
****** FORTRAN RUN-TIME SYSTEM ******
Error 1010: record too long
Location: the WRITE statement at line 8 of "test.f90"
Aborted
(It is interesting that Oracle Fortran reports the record to be "too long", which may refer to the input string.)