Please help a poor idiot with 'regular expressions'

[mdibella]

Here is my situation: I have a bunch of books whose author name has (as an example)

John Ringo - The Legacy of the Aldenata 5

What I want to do is move the stuff after the dash into the 'series' tag and the 'series index' tag, using 'Editing metadata in bulk' search and replace feature. Unless there's another way to do it, in which case I am *really* an idiot.

This expression: .*? -
returns the part I need to split into two pieces.

This expression: .*? - [^0-9]+
returns the number that goes into the index.


Can some genius please tell me how to construct an expression that will return the part between the dash and the number so I can put that into the series???

[Manichean]

I think you have misunderstood something. The expressions you gave don't "return" that part- I think you left the replace text empty and replaced everything but the part you want to get by an empty string. What you're looking for are backreferences, for example, any previously matched group (stuff in parentheses) can be referenced by its number of occurence: In your example, using the expression

(.*?) - (.*?)

with the replace string

\1

would return the author name, the replace string

\2

would return the series name and index.
What you'll need to do, then, is to do two passes: The first one inserts the series info into the series field, the second removes the series info from the author field. For the first one, use the search expression

.*? - (.*?) (\d+)

and the replace text

\1 \[\2\]

(Hint: This should, from what I remember, immediately set the series index as well- the stuff in the square brackets. Also, be careful to set your destination field to series.) For the second pass, use the search expression

(.*?) - .*?

and the replacement text

\1

(The destination field should now be authors)

Caveat: I didn't test this, so it may go horribly wrong. Test on a small sample size before going fullscale.
You may be interested in the backreferences part of the regex tutorial.

[mdibella]

Thanks so much! It didn't quite work but I figured it out.

The first pass has to use this as the replace text: \1 [\2]

[Manichean]

Of course, my brain was in regex mode. Sorry. Glad it worked.