Problems with format descriptors in Fortran - fortran

I'm learning Fortran and found some strange things when writing with a format (I'm using Fortran onlinegdb)
Program Hello
real, dimension(3,2):: array
array = 0
write(*, '(A,/, A,/, F5.2, F5.2)') &
"1","2",((array(i, j), i = 1,3), j = 1,2)
End Program Hello
I expected
1
2
0.00 0.00
0.00 0.00
0.00 0.00
I get
1
2
0.00 0.00
0.00 0.00
What's wrong?

Vladimir F is correct in saying that the format given does not suit the items that are provided for output: with format reversion after writing two real values, the control goes back to looking at the edit descriptor A but what corresponds to that isn't another character variable. This is not allowed.
However, the format suggested in an earlier revision of that other answer also does not give the output that you expect. If you want to write pairs of numbers on each line relying on an unlimited repeat specification, you'll need to explicitly put the file positioning into the format:
write(*, '(2(A,/),*(2F5.2,:,/))') "1", "2", transpose(array)
Without the / edit in there at the end, the repeat will mean that all elements of the array go in the same record. We also have : there, so that we don't get an extra line break after the final array element.
(I've also transposed the output array, as that's probably what you really mean. The implied do loop in the original output is a little unexpected and makes more sense moving over the final index first.)
With a limited repeat specification, as shown in the corrected form of that answer, format reversion does imply positioning:
write(*, '(2(A,/), 2(F5.2))') "1","2", transpose(array)
After processing the 2(F5.2) reversion has this reused while there are still elements to write out.
In summary, if you are relying on format reversion to "skip" earlier parts of a format while keeping new records, you must correctly mark the part of the whole format to revert to using parentheses. With just the whole format surrounded by parentheses, and no others, format reversion reuses the whole format.

You requested two strings to be printed, each on a new line, then two floating point numbers. That happened correctly.
But then there were still remaining items in the array. The format interpretation started again from the beginning with a new line and again for two strings and two numbers. But the array did not contain any strings...
Try '(A,/, A,/, (F5.2, F5.2))' instead. That will repeat the two-times floating point number format group until there are still numbers to be processed but this time the format does not return to the very beginning. (Note: the old and untested revision of the answer featured an extra repeat count - I did not realize this will disable the format reversion.)

Related

Controlling newlines when writing out arrays in Fortran

So I have some code that does essentially this:
REAL, DIMENSION(31) :: month_data
INTEGER :: no_days
no_days = get_no_days()
month_data = [fill array with some values]
WRITE(1000,*) (month_data(d), d=1,no_days)
So I have an array with values for each month, in a loop I fill the array with a certain number of values based on how many days there are in that month, then write out the results into a file.
It took me quite some time to wrap my head around the whole 'write out an array in one go' aspect of WRITE, but this seems to work.
However this way, it writes out the numbers in the array like this (example for January, so 31 values):
0.00000 10.0000 20.0000 30.0000 40.0000 50.0000 60.0000
70.0000 80.0000 90.0000 100.000 110.000 120.000 130.000
140.000 150.000 160.000 170.000 180.000 190.000 200.000
210.000 220.000 230.000 240.000 250.000 260.000 270.000
280.000 290.000 300.000
So it prefixes a lot of spaces (presumably to make columns line up even when there are larger values in the array), and it wraps lines to make it not exceed a certain width (I think 128 chars? not sure).
I don't really mind the extra spaces (although they inflate my file sizes considerably, so it would be nice to fix that too...) but the breaking-up-lines screws up my other tooling. I've tried reading several Fortran manuals, but while some of the mention 'output formatting', I have yet to find one that mentions newlines or columns.
So, how do I control how arrays are written out when using the syntax above in Fortran?
(also, while we're at it, how do I control the nr of decimal digits? I know these are all integer values so I'd like to leave out any decimals all together, but I can't change the data type to INTEGER in my code because of reasons).
You probably want something similar to
WRITE(1000,'(31(F6.0,1X))') (month_data(d), d=1,no_days)
Explanation:
The use of * as the format specification is called list directed I/O: it is easy to code, but you are giving away all control over the format to the processor. In order to control the format you need to provide explicit formatting, via a label to a FORMAT statement or via a character variable.
Use the F edit descriptor for real variables in decimal form. Their syntax is Fw.d, where w is the width of the field and d is the number of decimal places, including the decimal sign. F6.0 therefore means a field of 6 characters of width with no decimal places.
Spaces can be added with the X control edit descriptor.
Repetitions of edit descriptors can be indicated with the number of repetitions before a symbol.
Groups can be created with (...), and they can be repeated if preceded by a number of repetitions.
No more items are printed beyond the last provided variable, even if the format specifies how to print more items than the ones actually provided - so you can ask for 31 repetitions even if for some months you will only print data for 30 or 28 days.
Besides,
New lines could be added with the / control edit descriptor; e.g., if you wanted to print the data with 10 values per row, you could do
WRITE(1000,'(4(10(F6.0,:,1X),/))') (month_data(d), d=1,no_days)
Note the : control edit descriptor in this second example: it indicates that, if there are no more items to print, nothing else should be printed - not even spaces corresponding to control edit descriptors such as X or /. While it could have been used in the previous example, it is more relevant here, in order to ensure that, if no_days is a multiple of 10, there isn't an empty line after the 3 rows of data.
If you want to completely remove the decimal symbol, you would need to rather print the nearest integers using the nint intrinsic and the Iw (integer) descriptor:
WRITE(1000,'(31(I6,1X))') (nint(month_data(d)), d=1,no_days)

Meaning of 3F7.1 in Fortran data format

I am trying to create an MDM file using HLM 7 Student version, but since I don't have access to SPSS I am trying to import my data using ASCII input. As part of this process I am required to input the data format Fortran style. Try as I might I have not been able to understand this step. Could someone familiar with Fortran (or even better HLM itself) explain to me how this works? Here is my current understanding
From the example EG3.DAT they give
(A4,1X,3F7.1)
I think
A4 signifies that the ID is 4 characters long.
1X means skip a space.
F.1 means that it should read 1 decimal places.
I am very confused about what 3F7 might mean.
EG3.DAT
2020 380.0 40.3 12.5
2040 502.0 83.1 18.6
2180 777.0 96.6 44.4
Below are examples from the help documents.
Rules for format statement
Format statement example
EG1 data format
EG2 data format
EG3 data format
One similar question is Explaining Fortran Write Format. Unfortunately it does not explicitly treat the F descriptor.
3F7.1 means 3 floating point numbers, each printed over 7 characters, each with one decimal number behind the decimal point. Leading characters are blanks.
For reading you don't need the .1 info at all, just read a floating point number from those 7 characters.
You guessed the meaning of A4 (string of four characters) and 1X (one blank) correctly.
In Fortran, so-called data edit descriptors (which format the input or output of data) may have repeat specifications.
In the format (A4,1X,3F7.1) the data edit descriptors are A4 and F7.1. Only F7.1 has a repeat specification (the number before the F). This simply means that the format is as though the descriptor appeared repeated: like F7.1, F7.1, F7.1. With a repeat specification of 1, or not given, there is just the single appearance.
The format of the question, then, is like
(A4,1X,F7.1,F7.1,F7.1)
This format is one that is covered by the rules provided in one of the images of the question. In particular, the aspect of repeat specification is given in rule 2 with the corresponding example of rule 3.
Further, in Fortran proper, a repeat count specifier may also be * as special case: that's like an exceptionally large repeat count. *(F7.1) would be like F7.1, F7.1, F7.1, .... I see no indication that this is supported by HLM but if this is needed a very large repeat count may be given instead.
In 1X the 1 isn't a repeat specification but an integral, and necessary, part of the position edit descriptor.
Procedure for making MDM file from excel for HLM:
-Make sure ALL the characters in ALL the columns line up
Select a column, then right click and select Format Cells
Then click on 'Custom' and go to the 'Type' box and enter the number
of 0s you need to line everything up
-Remove all the tabs from the document and replace them with spaces.
Open the document in word and use find and replace
-To save the document as .dat
First save it as .txt
Then open it in Notepad and save it as .dat
To enter the data format (FORTRAN-Style)
The program wants to read the data file space by space, so you have to specify it perfectly so that it reads the whole set properly.
If something is off, even by a single space, then your descriptive stats will be wonky compared to if you check them in another program.
Enclose the code with brackets ()
Divide the entries with commas ,
-Need ID column for all levels
ID column needs to be sorted so that it is in order from smallest to
largest
Use A# with # being the number of characters in the ID
Use an X1 to
move from the ID to the next column
-Need to say how many characters are needed in each column
Use F
After F is the number of characters needed for that column -Use F# (#= number)
There need to be enough character spaces to provide one 'gap' space
between each column
There need to be enough to character spaces to allow for the decimal
As part of the F you need to specify the number of decimal places
You do this by adding a decimal point after the F number and then a
number to represent the spaces you need -F#.#
You can use a number in front of the F so as to 'repeat' it. Not
necessary though. -#F#.#
All in all, it should look something like this:
(A4,X1,F4.0,F5.1)
Helpful links:
https://books.google.de/books?id=VdmVtz6Wtc0C&pg=PA78&lpg=PA78&dq=data+format+fortran+style+hlm&source=bl&ots=kURJ6USN5e&sig=fdtsmTGSKFxn04wkxvRc2Vw1l5Q&hl=en&sa=X&ved=0ahUKEwi_yPurjYrYAhWIJuwKHa0uCuAQ6AEIPzAC#v=onepage&q&f=false
http://www.ssicentral.com/hlm/help6/error/Problems_creating_MDM_files.pdf
http://www.ssicentral.com/hlm/help7/faq/FAQ_Format_specifications_for_ASCII_data.pdf

Extract numbers out of text with inconcistant linebreaks

I have text with 6 numbers typically stored in one line
SomeData\n0.00 0.00 0.00 31,570.07 0.00 31,570.07\nSomeData
SomeData\n0.00 0.00 0.00 485,007.24 0.00 485,007.24\nSomeData
This regex worked fine on it:
\n[0-9,.-]* [0-9,.-]* [0-9,.-]* [0-9,.-]* [0-9,.-]* [0-9,.-]*\n
I noticed that every once in a while I get this:
SomeData\n0.00 0.00 10,921,594\n.89\n-\n9,563,271.0\n6\n0.00 1,358,323.83\nSomeData
Note how the linebreaks are randomly inserted after a sign or between numbers as if the system stored the values without filtering linebreaks.
I am struggling to get this extracted. I tried various expressions but my more successful one was [0-9,.-][\n]{0,1}[0-9,.-][ ]{0,1} to match an individual number.
What expression can I use to match both variations of the number formats preferably already stripping out the inconstant line breaks?
Update: Going with
[-\n]{0,2}[0-9,]+[\n.0-9]{3,4}[\n ]{0,1}
Please let me know if I there's a better way
One way would be to write an exact representation of what constitutes a number, so in your case [-+]?[0-9]+[0-9,]*(?:\.[0-9]+)? would do the trick. This helps, because then your search can know when a number starts and when one ends (because of rules like: a sign always is at the start a dot cannot appear multiple times, etc.). Then you want to match pairs of six delimited by either a new line or space so wrap it in a capture group and limit by 6: (...[ \n]*){6,6}. This helps because then the regex engine can figure out by backtracking what to consider a number by knowing how many it should match. Then you want to allow new lines in pretty much any position, so place the new line in each character group. You might also want to anchor the numbers on both sides, but this is not necessary, because now the regex engine will try to identify valid tuples of 6 numbers. End result is:
SomeData\n([-+]?[0-9\n]+[0-9,\n]*(?:\.[0-9\n]+)?[ \n]){6,6}SomeData
This will find tuples of 6 numbers no matter where the enters are. Here is an example: https://regex101.com/r/jD5nT8/1

Error when reading in float in Fortran

This should be quite simple, but I can't manage to read in a floating point number in Fortran. My program test.f looks like this:
PROGRAM TEST
open(UNIT=1,FILE='test.inp')
read(1,'(f3.0)')line
STOP
END
The input file test.inp simply contains a single float: 1.2
Now the compiling of my testfile goes fine, but when I run it I get an error:
At line 4 of file test.f (unit = 1, file = 'test.inp')
Fortran runtime error: Expected REAL for item 1 in formatted transfer, got INTEGER
(f3.0)
^
I've tried different modifications of the code and also googling for the error message, but with no result. Any help would be greatly appreciated!
Regards,
Frank
Your variable line is implicitly defined as integer. This doesn't work with thef edit descriptor. If you want to read an integer use i edit descriptor (i3 for example). Otherwise declare line as real to math the "f" descriptor.
Note beside: the .0 is not a problem, because if Fortran gets a number with decimal point the .0 part in the descriptor is ignored. It is only used when an number without a decimal is entered and then it uses the number behind the decimal point in the desciptor to add a decimal point into the right place. For with F8.5, 123456789 is read as 123.45678. More ont this here http://software.intel.com/sites/products/documentation/hpc/compilerpro/en-us/fortran/lin/compiler_f/lref_for/source_files/pghredf.htm .
In your read statement
read(1,'(f3.0)')line
the f3.0 tells tour program to read 3 digits with 0 digits after the decimal (this is what the n.m syntax means). So I presume that the program is just reading 1 from the file (not 1.2), which is an integer. Try replacing that line with something like
read(1,'(f3.1)')line
although, if the number in your file is likely to change and be larger than 9.9 or have more than one decimal place you should increase the field width to something larger than 3.
See the documentation of the read intrinsic and for data edit descriptors for more information on reading and writing in Fortran.
Edit: the format specifier, the second argument in quotes in your read statment, has the form fw.d, where f indicates that the data to read is a floating point number, w is the width of the field including all blanks and decimal points and d specifies the number of digits to the right of the decimal point.
I would suggest reading/writing list formatted data, unless you have a very strong reason to do otherwise. assuming that you're reading in from a file with just a single float or integer in a single line, like this
123.45
11
42
then this should do the reading
real*8 :: x,y,z
open(1,file=filename)
read(1,*)x
read(1,*)y
read(1,*)z
close(1)

How can I left-justify numerical output in fortran?

I am writing some simple output in fortran, but I want whitespace delimiters. If use the following statement, however:
format(A20,ES18.8,A12,ES18.8)
I get output like this:
p001t0000 3.49141273E+01obsgp_oden 1.00000000E+00
I would prefer this:
p001t0000 3.49141273E+01 obsgp_oden 1.00000000E+00
I tried using negative values for width (like in Python) but no dice. So, is there a way to left-justify the numbers?
Many thanks in advance!
There's not a particularly beautiful way. However, using an internal WRITE statement to convert the number to a text string (formerly done with an ENCODE statement), and then manipulating the text may do what you need.
Quoting http://rsusu1.rnd.runnet.ru/develop/fortran/prof77/node168.html
An internal file WRITE is typically
used to convert a numerical value to a
character string by using a suitable
format specification, for example:
CHARACTER*8 CVAL
RVALUE = 98.6
WRITE(CVAL, '(SP, F7.2)') RVALUE
The WRITE statement will fill the
character variable CVAL with the
characters ' +98.60 ' (note that there
is one blank at each end of the
number, the first because the number
is right-justified in the field of 7
characters, the second because the
record is padded out to the declared
length of 8 characters).
Once a number has been turned into a
character-string it can be processed
further in the various ways described
in section 7. This makes it possible,
for example, to write numbers
left-justified in a field, ...
This is easier with Fortran 95, but still not trivial. Write the number or other item to a string with a write statement (as in the first answer). Then use the Fortran 95 intrinsic "ADJUSTL" to left adjust the non-blank characters of the string.
And really un-elegant is my method (I program like a cave woman), after writing the simple Fortran write format (which is not LJ), I use a combination of Excel (csv) and ultraedit to remove the spaces effectively getting the desired LJ followed directly by commas (which I need for my specific import format to another software). BF
If what you really want is whitespace between output fields rather than left-justified numbers to leave whitespace you could simply use the X edit descriptor. For example
format(A20,4X,ES18.8,4X,A12,4X,ES18.8)
will insert 4 spaces between each field and the next. Note that the standard requires 1X for one space, some of the current compilers accept the non-standard X too.
!for left-justified float with 1 decimal.. the number to the right of the decimal is how many decimals are required. Write rounds to the desired decimal space automatically, rather than truncating.
write(*, ['(f0.1)']) RValue !or
write(*, '(f0.1)') RValue
!for left-justified integers..
write(*, ['(i0)']) intValue !or
write(*, '(i0)') RValue
*after feedback from Vladimir, retesting proved the command works with or without the array brackets