I think you need to check that: when I try your Regex in Expresso, or a simple app:
string input = @"I want a Regex to detect sources like (Lee & others, 2012) I want a Regex to detect sources like (Lee & others, 2012) I want a Regex to detect sources like (Lee & others, 2012)";
string output = Regex.Replace(input, @"\(.*?\,\s\d\d\d\d([a-d]?)\)", "%%");
Console.WriteLine(output);
I get exactly what I expect, and you ask for:
I want a Regex to detect sources like %% I want a Regex to detect sources like %% I want a Regex to detect sources like %%
I.e. it detects each bracket pair correctly (the "%%" are only there so you can see what was removed - it works fine with "" as well).
So have a look at your sample data, and try cutting it down until you have a minimum subset which still displays the problem - it may not be what you think it is!
"Would you please try on this:
and her colleagues
(2013) used carbon dating to measure the
growth of new neurons in the brains of
people aged 19 to 92. More than 50 years
ago, aboveground testing of nuclear bombs
released carbon-14, or C-14, into the atmosphere. Since these nuclear tests were
banned in 1963, levels of C-14 in the atmosphere have declined at a regular and
well-known rate. Measuring the amount of C-14 concentration in neurons therefore
provided a “time-stamp” for the neurons, allowing researchers to determine when
the neurons had been generated.
The carbon-14 signature was unmistakable: Spalding and her colleagues found
clear evidence of neurogenesis after birth and, in fact, throughout the lifespan. They
were also able to calculate the rate of neurogenesis in a specific region of the brain,
called the hippocampus, a brain region involved in learning and memory. It turned
out that an average of 1,400 new neurons were being generated each day. The rate
declined only slightly with age. Other regions of the brain, however, did not show
evidence of neurogenesis.
Research on neurogenesis in humans and animals has uncovered a number of
intriguing findings. It is now generally accepted that newborn neurons develop into
mature functioning neurons in at least two regions of the human brain—the hippocampus
and the olfactory bulb, responsible for odor perception (Lee & others, 2012).
These newly generated"
Well that works as well - it eliminates everything between the first open bracket and the first close bracket after a comma and a space - which is exactly what you told it to do!
One way to solve this is to use Balancing Groups - which I'm *not* going to explain :laugh: - and try and remove anything inside the brackets that ends with four digits:
\((?>\((?<c>)|[^()]+|\)(?<-c>))*(?(c)(?!))\d{4}\)
But ... a regex is probably the wrong approach - you may need a more complicated natural language processor or you will possibly miss some edge cases.
I'd suggest that you get a copy of
Expresso[
^] - it's free, and it examines and generates Regular expressions.
"I use Expresso, sorry, the regex you gave doesn't work at all..."
:doh: I missed that it captured the first one (2013) but not the second ... had teh digits in the wrong place :O
Try this:
\((?>\((?<c>)|[^()]*\d{4}|\)(?<-c>))*(?(c)(?!))\)Which should catch both cases. Sorry about that ...
"I'm sorry, I din't quite understand. Is it hard for you to do without explaining: adding the possibility of 1 letter after the year number? I once worked with Regex Matches and groups, but they're not needed now..."
Try:
\((?>\((?<c>)|[^()]*\d{4}[a-zA-Z]?|\)(?<-c>))*(?(c)(?!))\)
Submit an article or tip