Add two delimiters after every sixth comma

Question

I'm trying to expand on this question but can't figure out this issue:

Let's say I've got a file roll.txt:

echo "'123456789','987651234','129873645','213456789','987612345','543216789','432156789','876543291','213465789','542637819','123456','23456','22234','3456','7890543','34567891,'2345','567'" >> roll.txt

I can place a newline after every sixth comma with the following sed command:

sed 's/,/,\n/6; P; D' roll.txt

'123456789','987651234','129873645','213456789','987612345','543216789',
'432156789','876543291','213465789','542637819','123456','23456',
'22234','3456','7890543','34567891,'2345','567'

However, when I try to place two newlines after every sixth comma:

sed 's/,/,\n\n/6; P; D' roll.txt

'123456789','987651234','129873645','213456789','987612345','543216789',
'432156789','876543291','213465789','542637819','123456','23456',
'22234','3456','7890543','34567891,'2345','567'

I instead get two newlines after the sixth comma, and four newlines after the 12th comma. Why? And how can I get two newlines after every sixth comma?

Answer 1

As written in steeldriver's comment, in each cycle you add two lines, but print and remove only one. This would grow worse for longer sequences, with 3 and 7 and 15 empty lines ...

Thus, don't do the replacement, if your first line is empty:

sed '/^\n/!s/,/,\n\n/6; P; D'

Answer 2

With GNU awk for mult-char RS you could just define each record as being 6 non-commas-then-comma fields:

$ echo "'123456789','987651234','129873645','213456789','987612345','543216789','432156789','876543291','213465789','542637819','123456','23456','22234','3456','7890543','34567891,'2345','567'" |
awk -v RS='([^,]*,){0,6}' 'RT{print RT}'
'123456789','987651234','129873645','213456789','987612345','543216789',
'432156789','876543291','213465789','542637819','123456','23456',
'22234','3456','7890543','34567891,'2345',

and if you wanted to make sure every output line is 6 fields and only ends with , when the last field is empty so it's a valid CSV you could do:

$ echo "'123456789','987651234','129873645','213456789','987612345','543216789','432156789','876543291','213465789','542637819','123456','23456','22234','3456','7890543','34567891,'2345','567'" |
awk -v n=6 'BEGIN{RS="([^,]*,){0,"n"}"; FS=OFS=","} RT{$0=gensub(/,$/,"",1,RT); $n=$n; print}'
'123456789','987651234','129873645','213456789','987612345','543216789'
'432156789','876543291','213465789','542637819','123456','23456'
'22234','3456','7890543','34567891,'2345',

Answer 3

Using Raku (formerly known as Perl_6)

If you want to combine elements in Raku you can batch them together:

~$  raku -ne 'put join "\n", .split(",").batch(6).map: *.join(",");' roll.txt
'123456789','987651234','129873645','213456789','987612345','543216789'
'432156789','876543291','213465789','542637819','123456','23456'
'22234','3456','7890543','34567891,'2345','567'

So to get two-newlines between each batch, just join on \n\n:

~$  raku -ne 'put join "\n\n", .split(",").batch(6).map: *.join(",");' roll.txt
'123456789','987651234','129873645','213456789','987612345','543216789'
'432156789','876543291','213465789','542637819','123456','23456'
'22234','3456','7890543','34567891,'2345','567'

Raku's batch function is equivalent to Raku's rotor(..., :partial) call. If you want to drop incomplete sets of 6 elements at the end just call rotor().

Finally, sometimes splitting doesn't always give you the answer you're looking for. In that case you can try combing through the data to extract elements of interest. The code below gives the exact equivalent to the answers above, but may be conceptually simpler. The only difficulty is that ' apostrophe can mess up one-liner quoting, so the character can be declared using its Unicode name \c[APOSTROPHE]:

~$ raku -ne 'put join "\n\n", .comb(/ \c[APOSTROPHE] \d+ \c[APOSTROPHE] /).batch(6).map: *.join(",");'  roll.txt
'123456789','987651234','129873645','213456789','987612345','543216789'
'432156789','876543291','213465789','542637819','123456','23456'
'22234','3456','7890543','2345','567'

https://unix.stackexchange.com/a/611077/227738
https://docs.raku.org/language/regexes
https://raku.org

Answer 4

Try this:

echo "'123456789','987651234','129873645','213456789','987612345','543216789','432156789','876543291','213465789','542637819','123456','23456','22234','3456','7890543','34567891','2345','567'" > roll.txt
cat roll.txt | sed -E "s/([^,]*,){6}/\0\n\n/g"

You can change the number of items repeated between commas by replacing it with the 6.

Answer 5

Using awk:

$ awk -F, '{for (i=1;i<NF;i++) printf "%s", $i FS ((i%6==0) ? ORS ORS: "") }END{print $NF; print ""}' file
'123456789','987651234','129873645','213456789','987612345','543216789',
'432156789','876543291','213465789','542637819','123456','23456',
'22234','3456','7890543','34567891,'2345','567'