Mirror of forum.nim-lang.org

6403 :: Awk-Style text processing with Nim

• cmc (orginal) [2020-06-04T08:51:10+02:00] view original

  Nim is so flexible!

  I recently was doing some text processing and decided to try awk, which really does amazingly well in its domain, ie taking some tabular data and crunching it and putting out some kind of text format.

  I had tried to do it in Nim first but there was too much boilerplate- that is, until I realized you can bolt strformat onto stdtmpl without bothering to use a proc, and have something similarly concise to awk scripts. This script expects TSV input with two columns and left-aligns the first, right aligns the second, creating a markdown table.

  #? stdtmpl(emit="stdout.write &")
  # import strformat, streams, parsecsv
  # var x: CsvParser
  # open(x, newFileStream(stdin), "-", '\t')
  
  My Report
  =========
  
  {"":-<20} {"":-<20}
  # while x.readRow():
  {x.row[0]:<20} {x.row[1]:>20}
  # end
  {"":-<20} {"":-<20}
  Run

  
  foo\t1
  bar\t2
  fuz\t3
  buz}t4
  Run

---

• mratsim (orginal) [2020-06-04T11:19:12+02:00] view original

  Now that's a hack that is worth a blog post.

---

• rufusroflpunch (orginal) [2020-06-04T17:12:18+02:00] view original

  Well, that is clever!

---

• pietroppeter (orginal) [2020-06-04T18:05:00+02:00] view original

  It is also a very nice, self-contained and general enough example that could be added to source code filters documentation. All other examples start with a proc and are used through an include. This actually makes it clear that Source code filters can be standalone and used for scripting. It definitely improved my knowledge on what Source Code Filters are!

  Thanks for sharing!

---

• timothee (orginal) [2020-06-15T22:22:23+02:00] view original

  same remarks as https://github.com/nim-lang/Nim/issues/14637#issuecomment-644338428 the standard nim version is:

  • more readable
  • syntax highlights correctly
  • interacts correctly with the rest of the language
  • leads to code that's more reusable
  • is correctly indented instead of relying on features only invented to support SCF (end if|case|for|etc)
  • supports block comments

  etc...

  If you have boilerplate, just use templates+macros, no need to bake in a sublanguage into nim. SCF is a misfeature that needs to be deprecated.

  import strformat, streams, parsecsv
  var x: CsvParser
  open(x, newFileStream(stdin), "-", '\t')
  template p(a: string) = echo &(a)
  p """
  
  My Report
  =========
  
  {"":-<20} {"":-<20}"""
  while x.readRow():
    p """{x.row[0]:<20} {x.row[1]:>20}"""
  p """{"":-<20} {"":-<20}"""
  Run

---