Grep reverse character order

The Jedi's solution is slick, as usual, but I fear it won't work in your case, as that solution is built around searching for \u, meaning A Single Uppercase Letter, and that won't capture any Chinese characters. I suppose one could replace each instance of \u with something like [\x{2E80}-\x{9FBB}] which should capture most Chinese characters. (There are exceptions, like for full-width punctuation. If you wind up using a solution like this, there are easy ways to test to find out if all of your glyphs are being caught by the Unicode-value regex I provided above, like applying a highlight character style to visually review, which can be deleted from the document once you've done your QA.)

But, if you don't mind my asking... why? Why are bank names mirror-reversed in your document? I'd suspect that maybe they've accidentally have some right-to-left text setting applied, which could just be turned off. That'd be less risky than trying to reverse glyph order. 

… Supposing the Starting point is "ABCD" (in capitals) and the Arrival point is "DCBA"!

3 simple Regex!

In my sample, the biggest word is "EFGHIJKLMN" that has 10 chars. So you will need to play:

Regex "_Fridnge_1" once

Regex "_Fridnge_2" 10 times (until no match!)

Regex "_Fridnge_3" once

Of course, with a tool as Multi-Find/Change, you just need 1 click!

(^/)  The Jedi