Easy way to squeeze chars from many to one ?

[pebau]

Is there an easier way to reduce multiple Blanks to a single Blank Character inside a string ?
How could I write this in BaCon, without using EXEC$ and the tr command for squeezing ? 
Maybe its better to use linux.die.net/man/2/stat for getting the filesize ?

FOR file$ IN ls("-l")
    IF REGEX (file$,".bac$") THEN
      PRINT "A BaCon File:",file$
      '  echo "This    is for     testing" | tr -s [:space:]  " "    'squeeze via bash
      trim$="echo " & CHR$(34) & file$ & CHR$(34) & " | tr -s [:space:] " & CHR$(39) & " " & CHR$(39)
      tr$=EXEC$(trim$)
      PRINT "Filesize: ",TOKEN$(tr$,5," ")    ' in ls -l the filesize is on position 5
ENDIF
NEXT


FUNCTION ls (STRING param)
  RETURN EXEC$("ls" & " " & param)
END FUNCTION

[Pjot]

Jan 23, 2020 21:54:00 GMT 1 pebau said:

Is there an easier way to reduce multiple Blanks to a single Blank Character inside a string ?

Of course, that is very easy:

PRINT REPLACE$(a$, "[[:space:]]+", " ", TRUE)


Jan 23, 2020 21:54:00 GMT 1 pebau said:

Maybe its better to use linux.die.net/man/2/stat for getting the filesize ?

I don't see why? What is it you want to achieve?

BR
Peter

[pebau]

Hello, I just wanted to count files from an "ls -l" output with a certain file extension like *.bac or *.log
Then I had a working bash onliner with the bc calculator.

echo Bytes:; (ls -l ./*bac | awk '{print $5}' | tr "\012" "+"; echo 0 ) | bc

Then i thought lets do it in BaCon. After that i have seen that the TOKEN$ function
can not use a delimiter which could be "one blank or could be more blanks", like it is in ls -l outputs.

VAL(TOKEN$(file$,5," ")) is about the same like awk '{print $5}', but awk knows how to work
with different numbers of blanks (delimiters) and justs jumps over them in the input stream data flow
until it finds the next real data

So REPLACE$ with Regexes is needed to squeeze the Blanks, thank you Pjot !

[[:blank:]] means space and tab. This makes it similar to: [ \t].
[[:space:]], in addition to space and tab, includes newline, linefeed, formfeed, and vertical tab. This makes it similar to: [ \t\n\r\f\v].

Current Version works good ! (but still slower than bash)

GLOBAL sum_filesize TYPE int

FOR file$ IN ls("-l")
  IF REGEX (file$,".bac$") THEN    ' files which end with .bac
    sum_filesize=sum_filesize + VAL(TOKEN$(file$,5," ")) 
  ENDIF
NEXT
PRINT "Sum: ",sum_filesize


FUNCTION ls (STRING param)
    retval$=EXEC$("ls" & " " & param)
    retval$=REPLACE$(retval$,"[[:blank:]]+"," ",1)       ' keep Newlines, squeeze Blanks
      RETURN retval$
END FUNCTION

Python version: (about same performance as BaCon, slower than bash)

#!/usr/bin/python3.6
import os
my_dir="/home/peter/src_bacon/"
complete_size=0

for filename in os.listdir(my_dir):
    if filename.endswith(".bac"):
      complete_size+=(os.stat(my_dir+filename).st_size)   #get the filesize with os.stat
print ("Complete Size:",complete_size,"Bytes")


BR,
pebau

[Pjot]

How about using WALK$?

Count the files with ".bac" extension:

PRINT AMOUNT(WALK$("/path/to/my/dir", 1, ".bac$", FALSE, NL$), NL$)


Obtain the total file size:

path$ = "/path/to/my/dir/"
FOR f$ IN WALK$(path$, 1, ".bac$", FALSE, NL$) STEP NL$
   INCR total, FILELEN(path$ & f$)
NEXT
PRINT total


HTH
Peter

[pebau]

Thank You Pjot, i should have started with WALK$ and not "ls -l"  
Another awk version: 

ls -l *.bac | awk '{sum += $5} END {print sum}'

BR,
Pebau

Jan 24, 2020 15:18:25 GMT 1 Pjot said:

How about using WALK$?

Count the files with ".bac" extension:

PRINT AMOUNT(WALK$("/path/to/my/dir", 1, ".bac$", FALSE, NL$), NL$)


Obtain the total file size:

path$ = "/path/to/my/dir/"
FOR f$ IN WALK$(path$, 1, ".bac$", FALSE, NL$) STEP NL$
   INCR total, FILELEN(path$ & f$)
NEXT
PRINT total


HTH
Peter