格式化读取和暗示做 [英] Formatted reads and implied do

问题描述

在Fortran中，可以使用隐含循环。它们通常用于打印并具有以下结构。

  write（*，'（5I6）'）（i，i = 1,20）
 
！输出
 1 2 3 4 5 
 6 7 8 9 10 
 11 12 13 14 15 
 16 17 18 19 20 
  

这将允许写入5个宽度为6的整数，完成后自动进入下一行。因此，格式化非常有用。

问题是，如果这可以在相反的情况下完成

  read （*，'（5I6）'）（a（i），i = 1,20）
  

这当然不起作用，因为每行只有5个元素。从来就不是我想问如何在自动模式下翻到下一行。我目前正在统计入场次数，并且一旦达到5次，我就去下一行。另一种选择是在不存在新行的二进制暂存文件中读取文件。这对我来说不太合适，因为我用 index（）函数使用关键字搜索。

解决方案

我会这样做，如果你真的需要格式的原因

  do i = 1，n，5 
 read（*，'（5I6）'）a（i：i + 4）
 end do 
  

否则，按照高性能标记的建议，列表定向格式将运行良好。

您可以将其编写为隐含的，但是子数组更好。

我甚至可以想象一个带有格式描述符的单线程，包括 / ，但为什么要麻烦？

（太长的评论）

In Fortran one has the option to use implied loops. They are usually used for printing and have the following structure.

write(*,'(5I6)') (i,i=1,20)

! output 
 1     2     3     4     5
 6     7     8     9    10
11    12    13    14    15
16    17    18    19    20

This would allow to write 5 integers of width 6 and when done, go automatically to the next line. Therefore, it is quite useful for formatting.

The question is if this can be done in reverse

read(*,'(5I6)') (a(i),i=1,20)

This would of course not work, as it is written, because there are only 5 elements per line. Never the less I would like to ask how to flip to the next line in an automatic mode. I am currently counting the number of entries and once 5 is reached, I go to the next line. Another option would be to read the file in a binary scratch file, where no 'new' lines exist. This would not work neatly for me, because I use keywords search with the index() function.

解决方案

I would do it like this if you really need the format for some reason

do i = 1, n, 5
  read(*,'(5I6)') a(i:i+4)
end do

otherwise the list directed format would work well as suggested by High Performance Mark.

You could write it as an implied do, but sub-arrays are better.

I could even imagine a one liner with a format descriptor including /, but why bother?

(too long for a comment)

这篇关于格式化读取和暗示做的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！