Reputation: 1315
I am new to bash scripting, and I'm having an issue with one of my scripts. I'm trying to compose a list of Drivers Under 25 after reading their birthdates in from a folder filled with XML files and calculating their ages. Once I have determined they are under 25, the filename of the driver's data is saved to a text file. The script is working up until a certain point and then it stops. The error I'm getting is:
gdate: extra operand ‘+%s’
Try 'gdate --help' for more information.
DriversUnder25.sh: line 24: ( 1471392000 - )/60/60/24 : syntax error: operand expected (error token is ")/60/60/24 ")
Here is my code:
#!/bin/bash
# define directory to search and current date
DIRECTORY="/*.xml"
CURRENT_DATE=$(date '+%Y%m%d')
# loop over files in a directory
for FILE in $DIRECTORY;
do
# grab user's birth date from XML file
BIRTH_DATE=$(sed -n '/Birthdate/{s/.*<Birthdate>//;s/<\/Birthdate.*//;p;}' $FILE)
# calculate the difference between the current date
# and the user's birth date (seconds)
DIFFERENCE=$(( ( $(gdate -ud $CURRENT_DATE +'%s') - $(gdate -ud $BIRTH_DATE +'%s') )/60/60/24 ))
# calculate the number of years between
# the current date and the user's birth date
YEARS=$(($DIFFERENCE / 365))
# if the user is under 25
if [ "$YEARS" -le 25 ]; then
# save file name only
FILENAME=`basename $FILE`
# output filename to text file
echo $FILENAME >> DriversUnder25.txt
fi
done
I'm not sure why it correctly outputs the first 10 filenames and then stops. Any ideas why this may be happening?
Upvotes: 2
Views: 6233
Reputation: 1315
The issue was that there were multiple drivers in some files, thus importing multiple birth dates into the same string. My solution is below:
#!/bin/bash
# define directory to search and current date
DIRECTORY="/*.xml"
CURRENT_DATE=$(date '+%Y%m%d')
# loop over files in a directory
for FILE in $DIRECTORY;
do
# set flag for output to false initially
FLAG=false
# grab user's birth date from XML file
BIRTH_DATE=$(sed -n '/Birthdate/{s/.*<Birthdate>//;s/<\/Birthdate.*//;p;}' $FILE)
# loop through birth dates in file (there can be multiple drivers)
for BIRTHDAY in $BIRTH_DATE;
do
# calculate the difference between the current date
# and the user's birth date (seconds)
DIFFERENCE=$(( ( $(gdate -ud $CURRENT_DATE +'%s') - $(gdate -ud $BIRTHDAY +'%s') )/60/60/24))
# calculate the number of years between
# the current date and the user's birth date
YEARS=$(($DIFFERENCE / 365))
# if the user is under 25
if [ "$YEARS" -le 25 ]; then
# save file name only
FILENAME=`basename $FILE`
# set flag to true (driver is under 25 years of age)
FLAG=true
fi
done
# if there is a driver under 25 in the file
if $FLAG == true; then
# output filename to text file
echo $FILENAME >> DriversUnder25.txt
fi
done
Upvotes: 0
Reputation: 295279
A best-practices implementation would look something like this:
directory=/ # patch as appropriate
current_date_unix=$(date +%s)
for file in "$directory"/*.xml; do
while IFS= read -r birth_date; do
birth_date_unix=$(gdate -ud "$birth_date" +'%s')
difference=$(( ( current_date_unix - birth_date_unix ) / 60 / 60 / 24 ))
years=$(( difference / 365 ))
if (( years < 25 )); then
echo "${file%.*}"
fi
done < <(xmlstarlet sel -t -m '//Birthdate' -v . -n <"$file")
done >DriversUnder25.txt
If this script needs to be usable my folks who don't have xmlstarlet
installed, you can generate an XSLT template and then use xsltproc
(which is available out-of-the-box on modern opertaing systems).
That is to say, if you run this once, and bundle its output with your script:
xmlstarlet sel -C -t -m '//Birthdate' -v . -n >get-birthdays.xslt
...then the script can be modified to replace xmlstarlet
with:
xsltproc get-birthdays.xslt - <"$file"
Notes:
for file in "$directory"/*.xml
, the expansion is quoted but the glob is not (thus allowing the script to operate on directories with spaces, glob characters, etc. in their names).Upvotes: 1
Reputation: 530922
You need to quote the expansion of $BIRTH_DATE
to prevent word splitting on the whitespace in the value. (It is good practice to quote all your parameter expansions, unless you have a good reason not to, for this very reason.)
DIFFERENCE=$(( ( $(gdate -ud "$CURRENT_DATE" +'%s') - $(gdate -ud "$BIRTH_DATE" +'%s') )/60/60/24 ))
(Based on your comment, this would probably at least allow gdate
to give you a better error message.)
Upvotes: 3