Automatically round off - awk command - awk

Question

I have a csv file and I want to add a column that takes some values from other columns and make some calculations. As a simplified version I'm trying this:
awk -F"," '{print $0, $1+1}' myFile.csv |head -1
The output is:
29.325172701023977,...other columns..., 30
The column added should be 30.325172701023977 but the output is rounded off.
I tried some options using printf, CONVFMT and OFMT but nothing worked.
How can I avoid the round off?

Answer 1

Assumptions:
the number of decimal places is not known beforehand
the number of decimal places can vary from line to line
Setup:
$ cat myfile.csv
29.325172701023977,...other columns...
15.12345,...other columns...
120.666777888,...other columns...
46,...other columns...
One awk idea where we use the number of decimal places to dynamically generate the printf "%.?f" format:
awk '
BEGIN { FS=OFS="," }
{ split($1,arr,".") # split $1 on period
numdigits=length(arr[2]) # count number of decimal places
newNF=sprintf("%.*f",numdigits,$1+1) # calculate $1+1 and format with "numdigits" decimal places
print $0,newNF # print new line
}
' myfile.csv
NOTE: assumes OP's locale uses a decimal/period to separate integer from fraction; for a locale that uses a comma to separate integer from fraction it gets more complicated since it will be impossible to distinguish between a comma as integer/fraction delimiter vs field delimiter without some changes to the file's format
This generates:
29.325172701023977,...other columns...,30.325172701023977
15.12345,...other columns...,16.12345
120.666777888,...other columns...,121.666777888
46,...other columns...,47

Answer 2

as long as you aren't dealing with numbers greater than 9E15, there's no need to fudge any one of CONVFMT, OFMT, or s/printf() at all :
{m,g}awk '$++NF = int((_=$!__) + sub("^[^.]*",__,_))_' FS=',' OFS=','
29.325172701023977,...other columns...,30.325172701023977
15.12345,...other columns...,16.12345
120.666777888,...other columns...,121.666777888
46,...other columns…,47
if mawk-1 is sending ur numbers to scientific notation do :
mawk '$++NF=int((_=$!!NF)+sub("^[^.]*",__,_))_' FS=',' OFS=',' CONVFMT='%.f'
when u scroll right you'll notice all input digits beyond the decimal point are fully preserved
2929292929.32323232325151515151727272727270707070701010101010232323232397979797977,...other columns...,2929292930.32323232325151515151727272727270707070701010101010232323232397979797977
1515151515.121212121234343434345,...other columns...,1515151516.121212121234343434345
12121212120.66666666666767676767777777777788888888888,...other columns...,12121212121.66666666666767676767777777777788888888888
4646464646,...other columns…,4646464647
2929.32325151727270701010232397977,...other columns...,2930.32325151727270701010232397977
1515.121234345,...other columns...,1516.121234345
12120.66666767777788888,...other columns...,12121.66666767777788888
4646,...other columns...,4647
change it to CONVFMT='%\47.f', and you can even get mawk-1 to nicely comma format them for u :
29292929292929.323232323232325151515151515172727272727272707070707070701010101010101023232323232323979797979797977,...other columns…,29,292,929,292,930.323232323232325151515151515172727272727272707070707070701010101010101023232323232323979797979797977
15151515151515.12121212121212343434343434345,...other columns…,15,151,515,151,516.12121212121212343434343434345
121212121212120.666666666666666767676767676777777777777777888888888888888,…other columns…,121,212,121,212,121.666666666666666767676767676777777777777777888888888888888
46464646464646,...other columns...,46,464,646,464,647