sed

How to replace text in files?

# Replace text in original file: search for 'oldtext', replace with 'newtext', and save to same file.

sed -i 's/oldtext/newtext/g' myfile.txt

-i replace inside original file, not creating a new file

g replace all (globally): replace at all positions, not only the first where the search text appeared

# using variables: requires double quotes "s/oldtext/newtext/g"

FILENAME=myfile.txt

SEARCH_FOR=oldtext

REPLACE_WITH=newtext

sed -i "s/${SEARCH_FOR}/${REPLACE_WITH}/g" ${FILENAME}

# using a unix-pipe

cat myfile.txt | sed 's/oldtext/newtext/g' > newfile.txt

# replacing text without interpreting special characters

# Example: replace old long taxa-IDs with short taxa names in taxa_abundance_table.csv file

# get variables containing the search/replace text strings

OLDTAXA='k__Bacteria;p__Bacteroidetes;c__Bacteroidia;o__Bacteroidales;f__[Paraprevotellaceae];g__[Prevotella]'

NEWTAXA='Prevotella'

# mask/cover all special characters: especially replace [ with \[

OLDTAXA=$(sed -e 's/[]\/ ()$*.^|[]/\\&/g' <<< "${OLDTAXA}")

NEWTAXA=$(sed -e 's/[]\/ ()$*.^|[]/\\&/g' <<< "${NEWTAXA}")

echo ${OLDTAXA} ${NEWTAXA}

# replace taxa names in file: taxa_abundance_table.csv

sed -i "s/${OLDTAXA}/${NEWTAXA}/g" taxa_abundance_table.csv

# replace multiple taxa-IDs provided in a two column text file: taxa_old_new_names.txt

cat taxa_old_new_names.txt | while read OLDTAXA NEWTAXA; do

echo ${OLDTAXA} ${NEWTAXA} ;

OLDTAXA=$(sed -e 's/[]\/ ()$*.^|[]/\\&/g' <<< "${OLDTAXA}") ;

NEWTAXA=$(sed -e 's/[]\/ ()$*.^|[]/\\&/g' <<< "${NEWTAXA}") ;

sed -i "s/${OLDTAXA}/${NEWTAXA}/g" taxa_abundance_table.csv ;

done