Bdtyktl

I need to do a find and replace in all cards using the following settings:

As an example I want to find:

<span style="color: rgb(255, 255, 255)">Word</span>

and change it to

Word

In other words I want to strip those text around Word.

but the problem is I have other words too.like:

<span style="color: rgb(255, 255, 255)">One</span>

<span style="color: rgb(255, 255, 255)">Two</span>

<span style="color: rgb(255, 255, 255)">Three</span>

to

One

Two

Three

I want to change them all to the word between those tags.
How can I use the search and replace of the Anki to achieve that result by providing a regular expression?
What is the correct regular expression to achieve that result?

The answer seems easy to provide but when a valid regex is applied in the ‘Find and Replace’ it simply does not work.

The documentation https://apps.ankiweb.net/docs/manual.html#find-and-replace is not helpful at all and is even misleading as it shows an example which does not work when applied.

I can only assume that it was working in some of the previous versions but was unintentionally broken.


Let’s look at couple of trials with the ‘Find and Replace’ window. I am using the following settings.


Observation 1

Input: <

Find: <

Replace With:

Output: <


Summary: Nothing happened. The pattern is valid but it did not work as expected.


Observation 2

Input: <

Find: ^.

Replace With:

Output: lt;


Summary: lt; reminds html entity < Now we know that the regex engine is working but apparently the characters are stored differently than they look.



Anki stores collections in collection.anki2 file which is an sqlite database file. Let’s open it in DB Browser for SQLite and find the record with:
<span style="color: rgb(255, 255, 255)">One</span>
<span style="color: rgb(255, 255, 255)">Two</span>
<span style="color: rgb(255, 255, 255)">Three</span>

Column sfld:
<span style="color: rgb(255, 255, 255)">One<span><span style="color: rgb(255, 255, 255)">Two<span><span style="color: rgb(255, 255, 255)">Three<span>


Column flds:
<div>&lt;span style="color: rgb(255, 255, 255)"&gt;One&lt;span&gt;</div><div>&lt;span style="color: rgb(255, 255, 255)"&gt;Two&lt;span&gt;</div><div>&lt;span style="color: rgb(255, 255, 255)"&gt;Three&lt;span&gt;</div><div>&lt;span style="color: rgb(255, 255, 255)"&gt;One&lt;span&gt;</div><div>&lt;span style="color: rgb(255, 255, 255)"&gt;Two&lt;span&gt;</div><div>&lt;span style="color: rgb(255, 255, 255)"&gt;Three&lt;span&gt;</div>



Apparently regex works on the flds column.


Let’s try this regex:



Before change:



After change:



The pattern itself can be shortened but here I mainly focused on how to make it work at all.

I thought I would just chime in to say that @wlod's answer is correct.

You can create a capture group by surrounding any matching regex pattern you would like to keep during the replacement by surrounding it with parentheses and using the convention <digit> ordered by the number of (captured) groups you specified.

This killed me too before I learned of it because in many regex implementations, the $ is used to denote a capture group, but in Anki, the is used.

Example #1

You have various animal sentences in a field on your Anki notes.
You want to keep the animal and normalize the sentences for all of them.

# Card 1

Your dog is friendly.



# Card 2

That cat is evil.



# Card 3

My fish is dead.

Find:

.*(dog|cat|fish).*

Replace:

I love my 1.

Result:

# Card 1

I love my dog.



# Card 2

I love my cat.



# Card 3

I love my fish.

Example #2

Now let's say you also wanted to keep their describing adjectives. You will need two capture groups. One for the animal and one for the adjective.

You could do this.

# Card 1

Your dog is friendly.



# Card 2

That cat is evil.



# Card 3

My fish is dead.

Find:

.*(dog|cat|fish) is (.+).

Replace:

The 2 1

Result:

# Card 1

The friendly dog



# Card 2

The evil cat



# Card 3

The dead fish

I hope this helps clarify things.