在 Linux bash 中匹配文件名中的时间戳的 RegEx 问题

Question

I have the following code written to check if my filename ends with some format like given below. It is working when tested online at https://regex101.com/ but not in linux bash. Please help me find out the issue.

Ex: timestamp_foramt="^[\\s\\S]+_INFT_[0-9]{8}_[0-9]{6}[.][\\w]+$" FNAME="Payments.test_INFT_20191218_075918.txt" #example if [[ ! $FNAME =~ $timestamp_foramt ]]; then echo non-format; else echo format; fi non-format timestamp_foramt="^[\\s\\S]+_INFT_[0-9]{8}_[0-9]{6}[.][\\w]+$" FNAME="Payments.test_INFT_20191218_075918.txt" #example if [[ ! $FNAME =~ $timestamp_foramt ]]; then echo non-format; else echo format; fi non-format

Please help me understand why it is not working in bash. Did I commit something wrong in regex for timestamp_foramt ?

Answer 1

You can fix the current expression like this:

timestamp_foramt="^.+_INFT_[0-9]{8}_[0-9]{6}\.[[:alnum:]_]+$"
FNAME="Payments.test_INFT_20191218_075918.txt" #example
if [[ ! $FNAME =~ $timestamp_foramt ]]; then
  echo 'non-format';
else
  echo 'format';
fi;

See the online Bash demo .

Main points:

• [\\s\\S] can be used in Perl and other NFA regex engines to match any character, in POSIX ERE, used in Bash, you may simply use a .
• \\w is better written as [[:alnum:]_] in POSIX ERE as it is more portable to match letters, digits and underscores.