Regex examples

[ReaderRabbit]

Originally Posted by Steadyhands

You need to be in Regex mode in the search box for this to work.

I am in Mode: Regex and 'All HTML Files'

[eschwartz]

Originally Posted by Steadyhands

Note you will have to change the xxxx to whatever your paragraph style and there is a space after the \1

No, change the xxxx\d+ . (This may be why ReaderRabbit was no successful, as it certainly would lead to failure if the class does not have trailing numbers. Like in the example given: "indent".)

It woud be better to use a catchall like

Find:

((Mr|Mrs|Dr|other)\.)</p>\s*<p( [^>]*)?>

Replace:

\1

(space after the 1)

[DiapDealer]

Originally Posted by eschwartz

It woud be better to use a catchall like

Find:

((Mr|Mrs|Dr|other)\.)</p>\s*<p( [^>]*)?>

Replace:

\1

(space after the 1)

Which would exclude paragraphs that had no attributes. Why not:

((Mr|Mrs|Dr|other)\.)</p>\s*<p[^>]*?>

Replace:

\1

(space after the 1)
instead?

[eschwartz]

Originally Posted by DiapDealer

Which would exclude paragraphs that had no attributes. Why not:

((Mr|Mrs|Dr|other)\.)</p>\s*<p[^>]*?>

Replace:

\1

(space after the 1)
instead?

Reread my regex. I did think of that.

Nitpick: I used an optional capture group, which includes a space (because as a general idea, I like demanding spaces after the tag name and before the attribute, to avoid matching the wrong tags) -- but your regex does not need a ? because the star already covers that.

[ReaderRabbit]

Originally Posted by eschwartz

Reread my regex. I did think of that.

Nitpick: I used an optional capture group, which includes a space (because as a general idea, I like demanding spaces after the tag name and before the attribute, to avoid matching the wrong tags) -- but your regex does not need a ? because the star already covers that.

Thank you all for looking into this error for me.

[DiapDealer]

Originally Posted by eschwartz

Reread my regex. I did think of that.

My bad. I didn't notice the space was inside the optional grouping. Part of the reason I don't like to use a lot of extraneous grouping in my regex.

Originally Posted by eschwartz

Nitpick: I used an optional capture group, which includes a space (because as a general idea, I like demanding spaces after the tag name and before the attribute, to avoid matching the wrong tags)

If you're worried about matching the wrong tags (like <image for <i), just make sure there's a word break after the element name.

<p\b[^>]*>

Or if you're in calibre's editor, even the end-of-word match:

<p\M[^>]*>

[eschwartz]

Originally Posted by DiapDealer

My bad. I didn't notice the space was inside the optional grouping. Part of the reason I don't like to use a lot of extraneous grouping in my regex.


If you're worried about matching the wrong tags (like <image for <i), just make sure there's a word break after the element name.

<p\b[^>]*>

Or if you're in calibre's editor, even the end-of-word match:

<p\M[^>]*>

That works too.

Po-tay-to po-tah-to.

Doesn't confuse me, I got used to doing it this way (seems in my subjective opinion to make more sense), etc. Never really fell in love with word boundaries.

[DiapDealer]

Originally Posted by eschwartz

Po-tay-to po-tah-to.

Doesn't confuse me, I got used to doing it this way (seems in my subjective opinion to make more sense), etc. Never really fell in love with word boundaries.

Oh, absolutely. Lots of ways to skin the same cat. I don't love any of it, to tell the truth. I just find word boundaries to useful NOT to use. *shrug*

[JimmyG]

Okay this regex $(.+)^ finds whole lines with text in them.

But I don't want lines that start with <, so I tried this $([^<].+)^
but that includes any preceding blank line and the next line, whether it starts with < or not, and I don't know why.

I want whole lines (not empty) that don't start with a tag.

[mzmm]

Originally Posted by JimmyG

Okay this regex $(.+)^ finds whole lines with text in them.

But I don't want lines that start with <, so I tried this $([^<].+)^
but that includes any preceding blank line and the next line, whether it starts with < or not, and I don't know why.

I want whole lines (not empty) that don't start with a tag.

it might help if you post the code you're trying to parse and the output you expect, because the regex you're using doesn't really fit with what you're describing.

if you're trying to find lines in an html document that don't begin with a < (which would be very rare), then you could use something like

^\s*[^<\s]+

but i don't think that's really what you're after...

[DiapDealer]

Originally Posted by JimmyG

Okay this regex $(.+)^ finds whole lines with text in them.

But I don't want lines that start with <, so I tried this $([^<].+)^
but that includes any preceding blank line and the next line, whether it starts with < or not, and I don't know why.

I want whole lines (not empty) that don't start with a tag.

You got your anchors flipped ^ is the beginning of a string(or line) and $ is the end.

But even ^([^<].+)$ isn't going to be very useful in a Sigil formatted file. "Lines" get very hairy in a file. A paragraph is typically on one "line" (meaning no line-break characters) from <p> to </p>. Same with just about any block-level element. And many lines are likely to be indented, so they don't start with "<" they start with a space. That's probably why they're getting included in your search.

It's including blank lines because blank lines DON'T start with a "<", they start with line-break character(s).

There really shouldn't be any (or very, very few anyway) "lines" that don't begin with a "<" (or an indent before a "<"). Some css styling in the header and the like maybe.

If it's these relatively rare instances you're looking for perhaps something like:

^\w.+$

might come close?

[JimmyG]

Originally Posted by DiapDealer

You got your anchors flipped ^ is the beginning of a string(or line) and $ is the end.

But even ^([^<].+)$ isn't going to be very useful in a Sigil formatted file. "Lines" get very hairy in a file. A paragraph is typically on one "line" (meaning no line-break characters) from <p> to </p>. Same with just about any block-level element. And many lines are likely to be indented, so they don't start with "<" they start with a space. That's probably why they're getting included in your search.

It's including blank lines because blank lines DON'T start with a "<", they start with line-break character(s).

There really shouldn't be any (or very, very few anyway) "lines" that don't begin with a "<" (or an indent before a "<"). Some css styling in the header and the like maybe.

If it's these relatively rare instances you're looking for perhaps something like:

^\w.+$

might come close?

Yeah, I caught that $^. I was making my problem too complicated by trying to do it in Sigil. The lines without p tags are all contiguous, so I just copy them to EditPadPro (where ^([^<].*)$ does not include the blank line) and then copy them back to Sigil. Actually, doing it that way, ^(.+)$ should work, but the original worked so I just left it.

[Doitsu]

off-topic: I saw this Regex T-Shirt yesterday in the subway and it took me at least 5 minutes to figure it out.

[canpolat]

I think it should be:

(bb|[^2]b)

Currently it reads: two b or two not b.

[DiapDealer]

Be be or not be twice?