How to find a text between two characters and repl

Question

I have many lines of bibliography as the following:

@article{mehri_buckling_2016, title = {Buckling }, doi = {10.1016/j.cma.2016.01.017}}

@article{jin_modified_2014, title = {A modified ns}, doi = {10.1016/j.apacoust.2014.04.007}}

@article{sofiyev_free_2017, title = {The free}, doi = {10.1016/j.compositesb.2017.03.054}}

@article{malekzadeh_three-dimensional_2012, title = {Three-dimensional free }}

@article{jooybar_thermal_2016, title = {Thermal panels}, doi = {10.1016/j.tws.2016.01.032}}

In each line may there is doi={...} or not. I need to replace the words between @article{......, by the word between doi = {...}, if there is doi in its line.

For this example results should become:

@article{10.1016/j.cma.2016.01.017, title = {Buckling }, doi = {10.1016/j.cma.2016.01.017}}

@article{10.1016/j.apacoust.2014.04.007, title = {A modified ns}, doi = {10.1016/j.apacoust.2014.04.007}}

@article{10.1016/j.compositesb.2017.03.054, title = {The free}, doi = {10.1016/j.compositesb.2017.03.054}}

@article{malekzadeh_three-dimensional_2012, title = {Three-dimensional free }}

@article{10.1016/j.tws.2016.01.032, title = {Thermal panels}, doi = {10.1016/j.tws.2016.01.032}}

Answer 1

You could do it like this:

• Find what: ^(@article\{).*?(,.*?doi = \{)(.*?)\}
• Replace with: \1\3\2\3}
• Regular expression: checked
• . matches newline: cleared

Explanation:

• ^ is maybe optional, but it ensures that the match can only occur at the beginning of a line.
• (@article\{) matches literally @article{. The brace had to be escaped with a backslash, as it is a special character in regexes. The parentheses make it possible to refer to this string with \1 later. It's a so called "capture group"
• .*? will match any series of characters , but not more than necessary to get a complete match ("lazy"): . means any character, * any number of them, ? but as few as possible to make the rest work.
• (,.*?doi = \{) matches a literal comma, then again some more characters (see previous point), followed by a literal doi = {. The parentheses make it possible to refer to this as \2 later.
• (.*?) matches a number of characters (see above). The parentheses make it possible to refer to this as \3 later. This is the part that matches the actual text that needs to be copied.
• \} a literal }.

The replacement uses the capture groups:

• \1 is synonymous for @article{
• \3 is the text that needs to be copied
• \2 is whatever appeared between , and doi = {, including those two texts.
• \3 is repeated to just restore the doi value as it was.
• } is the closing brace that was matched at the end of the "find" pattern. We want to keep it as well of course.