print columns from bash variable

Hi, I have this bash variable: MARCA_TABELA="mencionadoo $7 $3 $1" and I want to add an extra column to a data file like this: cat $FICHEIRO_TEMP | awk --field-separator=";" --assign mt="${MARCA_TABELA}" '{OFS=";"; print $0 , mt }' > $FICHEIRO The problem is that awk does not add "mencionadoo content of column 7 "content of column 3 content of column 1" as the last column. Instead, for each line, awk adds: "mencionadoo $7 $3 $1". The expansion (or conversion) of $7 $3 $1 is not performed. How can I make it work? Luis

Luis P. Mendes wrote: > Hi, > > I have this bash variable: > MARCA_TABELA="mencionadoo $7 $3 $1" > > and I want to add an extra column to a data file like this: > cat $FICHEIRO_TEMP | awk --field-separator=";" --assign > mt="${MARCA_TABELA}" '{OFS=";"; print $0 , mt }' > $FICHEIRO Since it does not change, you should assign OFS in the BEGIN block, or using --assign (or its standard equivalent -v). > The problem is that awk does not add "mencionadoo content of column 7 > "content of column 3 content of column 1" > as the last column. > > Instead, for each line, awk adds: "mencionadoo $7 $3 $1". > The expansion (or conversion) of $7 $3 $1 is not performed. Awk does not do that kind of reevaluation. One solution here could be to create two shell variables in the first place, one with "mencionadoo" (and all other needed fixed strings), and another one with the list of fields to print, without the $, like eg "7 3 1". Then you could split the second variable and put the field numbers in an array. Example: $ echo "$FIXED" mencionadoo $ echo "$MARCA_TABELA" 7 3 1 $ awk -F ';' -v fixed="$FIXED" -v mt="$MARCA_TABELA" -v OFS=';' ' BEGIN{n=split(mt,ls,/ /)} {str=ofs="";for(i=1;i<=n;i++){str=str ofs $(ls[i]);ofs=OFS} print $0, fixed, str}' Or if you only have a single fixed part, you can do all the parsing in awk (assumes the first word is the fixed string): $ echo "$MARCA_TABELA" mencionadoo 7 3 1 $ awk -F ';' -v mt="$MARCA_TABELA" -v OFS=';' ' BEGIN{ match(mt,/ /); # look where first word ends fixed=substr(mt,1,RSTART-1) # extract fixed part mt=substr(mt,RSTART+1) # remove fixed part n=split(mt,ls,/ /) # get fields } {# main body same as before}' Hope that helps. -- awk 'BEGIN{O="~"~"~";o="=="=="==";o+=+o;x=O""O;while(X++<=x+o+o)c=c"%c"; printf c,(x-O)*(x-O),x*(x-o)-o,x*(x-O)+x-O-o,+x*(x-O)-x+o,X*(o*o+O)+x-O, X*(X-x)-o*o,(x+X)*o*o+o,x*(X-x)-O-O,x-O+(O+o+X+x)*(o+O),X*X-X*(x-O)-x+O, O+X*(o*(o+O)+O),+x+O+X*o,x*(x-o),(o+X+x)*o*o-(x-O-O),O+(X-x)*(X+O),x-O}'

Mon, 09 Mar 2009 13:50:36 +0000, Dave B escreveu: > Luis P. Mendes wrote: >> Hi, >> >> I have this bash variable: >> MARCA_TABELA="mencionadoo $7 $3 $1" >> >> and I want to add an extra column to a data file like this: cat >> $FICHEIRO_TEMP | awk --field-separator=";" --assign >> mt="${MARCA_TABELA}" '{OFS=";"; print $0 , mt }' > $FICHEIRO > > Since it does not change, you should assign OFS in the BEGIN block, or > using --assign (or its standard equivalent -v). > >> The problem is that awk does not add "mencionadoo content of column 7 >> "content of column 3 content of column 1" as the last column. >> >> Instead, for each line, awk adds: "mencionadoo $7 $3 $1". The expansion >> (or conversion) of $7 $3 $1 is not performed. > > Awk does not do that kind of reevaluation. One solution here could be to > create two shell variables in the first place, one with "mencionadoo" > (and all other needed fixed strings), and another one with the list of > fields to print, without the $, like eg "7 3 1". Then you could split > the second variable and put the field numbers in an array. Example: > > $ echo "$FIXED" > mencionadoo > $ echo "$MARCA_TABELA" > 7 3 1 > $ awk -F ';' -v fixed="$FIXED" -v mt="$MARCA_TABELA" -v OFS=';' ' > BEGIN{n=split(mt,ls,/ /)} > {str=ofs="";for(i=1;i<=n;i++){str=str ofs $(ls[i]);ofs=OFS} > print $0, fixed, str}' > > > Or if you only have a single fixed part, you can do all the parsing in > awk (assumes the first word is the fixed string): > > $ echo "$MARCA_TABELA" > mencionadoo 7 3 1 > $ awk -F ';' -v mt="$MARCA_TABELA" -v OFS=';' ' BEGIN{ > match(mt,/ /); # look where first word ends > fixed=substr(mt,1,RSTART-1) # extract fixed part > mt=substr(mt,RSTART+1) # remove fixed part n=split(mt,ls,/ /) > # get fields > } > {# main body same as before}' > > Hope that helps. Hello Dave, thank you for your support, but I still can't have it working. Please consider: $ cat xx.csv "a_col1";"a_col2";"a_col3";"a_col4";"a_col5";"a_col6";"a_col7";"a_col8";"a_col9";"a_col10";"a_col11" "b_col1";"b_col2";"b_col3";"b_col4";"b_col5";"b_col6";"b_col7";"b_col8";"b_col9";"b_col10";"b_col11" and the following script: 1 MARCA_TABELA="mencionadoo 7 3 1 " 2 echo "$MARCA_TABELA" 3 cp 1.csv xx.csv 4 cat xx.csv | awk -F ';' -v mt="$MARCA_TABELA" -v OFS=';' ' BEGIN{ 5 match(mt,/ /); # look where first word ends 6 fixed=substr(mt,1,RSTART-1) # extract fixed part 7 mt=substr(mt,RSTART+1) # remove fixed part n=split (mt,ls,/ /) 8 n=split(mt,ls,/ /) 9 } 10 {str=ofs="";for(i=1;i<=n;i++){str=str ofs $(ls[i]);ofs=OFS} 11 print str }' > yy.csv $ cat yy.csv "a_col7";"a_col3";"a_col1";"a_col1";"a_col2";"a_col3";"a_col4";"a_col5";"a_col6";"a_col7";"a_col8";"a_col9";"a_col10";"a_col11" "b_col7";"b_col3";"b_col1";"b_col1";"b_col2";"b_col3";"b_col4";"b_col5";"b_col6";"b_col7";"b_col8";"b_col9";"b_col10";"b_col11" It shows the 7th, 3rd and 1st columns as I'd expect, but repeats all the columns afterwards, and I don't understand why. Luis

Luis P. Mendes wrote: > Please consider: > $ cat xx.csv > "a_col1";"a_col2";"a_col3";"a_col4";"a_col5";"a_col6";"a_col7";"a_col8";"a_col9";"a_col10";"a_col11" > "b_col1";"b_col2";"b_col3";"b_col4";"b_col5";"b_col6";"b_col7";"b_col8";"b_col9";"b_col10";"b_col11" > > and the following script: > 1 MARCA_TABELA="mencionadoo 7 3 1 " > 2 echo "$MARCA_TABELA" > 3 cp 1.csv xx.csv > 4 cat xx.csv | awk -F ';' -v mt="$MARCA_TABELA" -v OFS=';' 'BEGIN{ > 5 match(mt,/ /); # look where first word ends > 6 fixed=substr(mt,1,RSTART-1) # extract fixed part > 7 mt=substr(mt,RSTART+1) # remove fixed part > 8 n=split(mt,ls,/ /) > 9 } > 10 {str=ofs="";for(i=1;i<=n;i++){str=str ofs $(ls[i]);ofs=OFS} > 11 print str }' > yy.csv > > $ cat yy.csv > "a_col7";"a_col3";"a_col1";"a_col1";"a_col2";"a_col3";"a_col4";"a_col5";"a_col6";"a_col7";"a_col8";"a_col9";"a_col10";"a_col11" > "b_col7";"b_col3";"b_col1";"b_col1";"b_col2";"b_col3";"b_col4";"b_col5";"b_col6";"b_col7";"b_col8";"b_col9";"b_col10";"b_col11" > > It shows the 7th, 3rd and 1st columns as I'd expect, but repeats all the > columns afterwards, and I don't understand why. Works fine for me. Are you sure you pasted the exact code you are using? Just copy/paste it, without adding line numbers or anything. -- awk 'BEGIN{O="~"~"~";o="=="=="==";o+=+o;x=O""O;while(X++<=x+o+o)c=c"%c"; printf c,(x-O)*(x-O),x*(x-o)-o,x*(x-O)+x-O-o,+x*(x-O)-x+o,X*(o*o+O)+x-O, X*(X-x)-o*o,(x+X)*o*o+o,x*(X-x)-O-O,x-O+(O+o+X+x)*(o+O),X*X-X*(x-O)-x+O, O+X*(o*(o+O)+O),+x+O+X*o,x*(x-o),(o+X+x)*o*o-(x-O-O),O+(X-x)*(X+O),x-O}'

Mon, 09 Mar 2009 17:17:59 +0000, Dave B escreveu: > Luis P. Mendes wrote: > >> Please consider: >> $ cat xx.csv >> "a_col1";"a_col2";"a_col3";"a_col4";"a_col5";"a_col6";"a_col7";"a_col8";"a_col9";"a_col10";"a_col11" >> "b_col1";"b_col2";"b_col3";"b_col4";"b_col5";"b_col6";"b_col7";"b_col8";"b_col9";"b_col10";"b_col11" >> >> and the following script: >> 1 MARCA_TABELA="mencionadoo 7 3 1 " >> 2 echo "$MARCA_TABELA" >> 3 cp 1.csv xx.csv >> 4 cat xx.csv | awk -F ';' -v mt="$MARCA_TABELA" -v OFS=';' 'BEGIN{ 5 >> match(mt,/ /); # look where first word ends 6 >> fixed=substr(mt,1,RSTART-1) # extract fixed part 7 >> mt=substr(mt,RSTART+1) # remove fixed part 8 >> n=split(mt,ls,/ /) >> 9 } >> 10 {str=ofs="";for(i=1;i<=n;i++){str=str ofs $(ls[i]);ofs=OFS} 11 >> print str }' > yy.csv >> >> $ cat yy.csv >> "a_col7";"a_col3";"a_col1";"a_col1";"a_col2";"a_col3";"a_col4";"a_col5";"a_col6";"a_col7";"a_col8";"a_col9";"a_col10";"a_col11" >> "b_col7";"b_col3";"b_col1";"b_col1";"b_col2";"b_col3";"b_col4";"b_col5";"b_col6";"b_col7";"b_col8";"b_col9";"b_col10";"b_col11" >> >> It shows the 7th, 3rd and 1st columns as I'd expect, but repeats all >> the columns afterwards, and I don't understand why. > > Works fine for me. Are you sure you pasted the exact code you are using? > Just copy/paste it, without adding line numbers or anything. Ok, there's the copy of it. MARCA_TABELA="mencionadoo 7 3 1 " echo "$MARCA_TABELA" cp 1.csv xx.csv cat xx.csv | awk -F ';' -v mt="$MARCA_TABELA" -v OFS=';' ' BEGIN{ match(mt,/ /); # look where first word ends fixed=substr(mt,1,RSTART-1) # extract fixed part mt=substr(mt,RSTART+1) # remove fixed part n=split (mt,ls,/ /) n=split(mt,ls,/ /) } {str=ofs="";for(i=1;i<=n;i++){str=str ofs $(ls[i]);ofs=OFS} print str }' > yy.csv I tried on the command line and the result is the same: $ cat yy.csv "a_col7";"a_col3";"a_col1";"a_col1";"a_col2";"a_col3";"a_col4";"a_col5";"a_col6";"a_col7";"a_col8";"a_col9";"a_col10";"a_col11" "b_col7";"b_col3";"b_col1";"b_col1";"b_col2";"b_col3";"b_col4";"b_col5";"b_col6";"b_col7";"b_col8";"b_col9";"b_col10";"b_col11" $ awk --version GNU Awk 3.1.5 Luis

Luis P. Mendes wrote: > Ok, there's the copy of it. > > MARCA_TABELA="mencionadoo 7 3 1 " Ah, here is the problem (it was there before too, sorry for missing it). Remove the trailing space: MARCA_TABELA="mencionadoo 7 3 1" The trailing space causes an empty element to be created in the array ls, which evaluates to $0 (ie, the whole line) when referenced with $(ls[i]). -- awk 'BEGIN{O="~"~"~";o="=="=="==";o+=+o;x=O""O;while(X++<=x+o+o)c=c"%c"; printf c,(x-O)*(x-O),x*(x-o)-o,x*(x-O)+x-O-o,+x*(x-O)-x+o,X*(o*o+O)+x-O, X*(X-x)-o*o,(x+X)*o*o+o,x*(X-x)-O-O,x-O+(O+o+X+x)*(o+O),X*X-X*(x-O)-x+O, O+X*(o*(o+O)+O),+x+O+X*o,x*(x-o),(o+X+x)*o*o-(x-O-O),O+(X-x)*(X+O),x-O}'

Mon, 09 Mar 2009 17:31:24 +0000, Dave B escreveu: > Luis P. Mendes wrote: > >> Ok, there's the copy of it. >> >> MARCA_TABELA="mencionadoo 7 3 1 " > > Ah, here is the problem (it was there before too, sorry for missing it). > Remove the trailing space: > > MARCA_TABELA="mencionadoo 7 3 1" > > The trailing space causes an empty element to be created in the array > ls, which evaluates to $0 (ie, the whole line) when referenced with > $(ls[i]). I printed n before and saw that its value was 4 instead of 3, but didn't guess the empty element would evaluate to $0. Thank you Dave, it works perfectly, now! Luis

Similiar Articles:

print columns from bash variable - comp.lang.awk
Hi, I have this bash variable: MARCA_TABELA="mencionadoo $7 $3 $1" and I want to add an extra column to a data file like this: cat $FICHEIRO_TEMP |...

passing awk variable to shell variable - comp.lang.awk
print columns from bash variable - comp.lang.awk Environment variables in an awk script - comp.lang.awk print columns from bash variable - comp.lang.awk passing awk ...

Newbie: group and print by column 1, awk? - comp.unix.shell ...
print columns from bash variable - comp.lang.awk print columns from bash variable - comp.lang.awk | Computer Group Hi ... bash, print variable column: bash, print, awk ...

read file with variable column numbers - comp.soft-sys.matlab ...
print columns from bash variable - comp.lang.awk... variable - comp.lang.awk Reading ASCII text file with variable number of ... print the first column in file. in a bash ...

Reading ASCII text file with variable number of columns - comp ...
print columns from bash variable - comp.lang.awk Reading ASCII text file with variable number of columns - comp ... print columns from bash variable - comp.lang.awk ...

Environment variables in an awk script - comp.lang.awk
print columns from bash variable - comp.lang.awk passing awk variable to shell variable - comp.lang.awk ... awk Passing shell vars to awk - comp.lang.awk print columns ...

AWK question - split string into variables - comp.unix.shell ...
> Here's what I mean by that: $ gawk 'BEGIN{ print mktime("2010 02 28 0 0 ... AWK question - split string into variables - comp.unix.shell ... split bash variable by ...

change field separator - comp.lang.awk
print columns from bash variable - comp.lang.awk change field separator - comp.lang.awk print all fields but $NF - comp.lang.awk change field separator - comp.lang.awk ...

print all fields but $NF - comp.lang.awk
print columns from bash variable - comp.lang.awk change field separator - comp.lang.awk print all fields but $NF - comp.lang.awk change field separator - comp.lang.awk ...

(sh/bash) How to check for a string matching -*? - comp.unix.shell ...
print columns from bash variable - comp.lang.awk (sh/bash) How to check for a string matching -*? - comp.unix.shell ... print columns from bash variable - comp.lang.awk ...

print columns from bash variable - comp.lang.awk | Computer Group
Hi, I have this bash variable: MARCA_TABELA="mencionadoo $7 $3 $1" and I want to add an extra column to a data file like this: cat $FICHEIRO_TEMP |...

use awk/gawk in bash, print variable column: bash, print, awk ...
awk '{ print $1 '} file will print the first column in file. in a bash script, I have a variable which contains column number: var=10; How do I refer this ...

7/22/2012 5:01:22 AM