13

Possible Duplicate:
How to “Copy as Unicode” from a Notebook?

How can I copy Greek text from notebooks as proper Unicode that can be pasted into other applications? If I type some Greek text into a notebook, and then try to paste it elsewhere, it ends up as a set of character names in Mathematica's notation. The screenshot below illustrated the problem.

Mathematica graphics

It doesn't really matter if the Greek text is typed elsewhere, then pasted into Mathematica, or if it's typed directly into Mathematica: when copying from Mathematica, Greek letters are converted into character names.

If you don't have a Greek keyboard layout installed, this can be useful for testing.

Use case: analysing Greek text and copying results from a notebook.

Community
  • 1
Szabolcs
  • 234,956
  • 30
  • 623
  • 1,263
  • I second @R.M.'s comment that it's an exact duplicate. But if we ever get a set of methods that works fo all platforms, it would be great to include that in the StackExchange Copy palette so that code with things like \[Element] becomes easier to read. – Jens May 31 '12 at 16:27
  • @R.M I edited the title. Now it's not a duplicate any more as it only refers to Greek. – Szabolcs May 31 '12 at 16:28
  • I don't think that changing the title makes this less of a duplicate. It's now just a special case. – Jens May 31 '12 at 16:30
  • @Jens It's not a duplicate because the answer I gave only applies to Greek text (my current use case) and not to other characters, such as \[LongRightArrow]. This answer will work for as long as the only concern is handling Greek text. But if the questions get merged, I don't mind. – Szabolcs May 31 '12 at 16:32
  • I just saw your answer - didn't think there would be such a specialized case, but I guess with that point of view one can leave this open as a separate issue. – Jens May 31 '12 at 16:33
  • @Jens I would have answered on the other thread, but the workaround is for Windows only, and has some potentially undesirable side effect. I wouldn't recommend it for those who only want to copy Greek variable names. However, for Greek text processing it's a useful workaround. – Szabolcs May 31 '12 at 16:34
  • 1
    I use vim to insta-replace all my \[Greek] but if you're using notepad, I have no help. – rm -rf May 31 '12 at 16:44
  • 1
    Looking at this again, I still cannot understand why you would not simply use Mr. Wizard's copy function from the related answer here. It seems so much more powerful!? – Jens Jun 01 '12 at 03:15
  • 1
    I feel a bit odd closing one of your questions in "favor" of my own, but I think this is a legitimate duplicate. I hope you agree; if not, please make the case to reopen. – Mr.Wizard Jun 01 '12 at 07:03
  • Hope this answer can help. – Silvia Mar 31 '14 at 06:31

2 Answers2

4

The following workaround works at least for Linux with Mathematica version 8.0.0.0; it might work on other systems, too:

Step 1: Open a terminal window (with UTF8 encoding) and start directly the kernel there.

Step 2: Enter $CharacterEncoding="UTF8" and press Enter. Ideally the kernel would figure that out itself from the locale, but for some reason it doesn't. Note: If you make a mistake here (e.g. forgetting the quotes, or mistyping the encoding) and get an error message about a non-supported encoding, quit the kernel in the terminal and start a new one. At least for 8.0.0.0, after an error, further assignments of that variable don't affect the output any more (I have no idea why).

Step 3: Copy/paste your output from the notebook to the terminal and press Enter. While the input is still in Mathematica \[...] form, the output will be in Unicode (be aware however that Mathematica uses private-use characters for some special characters like \[Equal]; those will not appear correctly — however that this is only relevant for characters in strings; Equal expressions will show up with ==. If all you need are Greek characters, you won't have any problems.)

Step 4: Copy/paste the output of step 3 to where you want it.

celtschk
  • 19,133
  • 1
  • 51
  • 106
3

I found the following workaround for Windows systems. Go to the Control Panel and select Regional & Language Options. Under the Advanced tab select "Greek" as the language for non-Unicode programs. Note that the option to change is not the language for "Standards and Formats". It is the language for non-Unicode programs.

Now Greek letters will be copied as Unicode from notebooks.

Warning: This might break the display of some characters in non-Unicode programs, and I won't be surprised if it has other side effects as well (e.g. non-Unicode programs may not be able to read from directories with certain accented characters in the name). Use with caution!

Szabolcs
  • 234,956
  • 30
  • 623
  • 1,263
  • With that many side-effects, I'm not sure if it'll be helpful... – rm -rf May 31 '12 at 16:42
  • @R.M It depends. This is the default on Greek versions of Windows and is probably useful as a default for Greek speakers who might name their directories in Greek. If you're going to do some text processing on Greek texts, then I think it's a useful workaround. If you want to copy variable names, then I wouldn't recommend it. (This is why I didn't simply post this answer on the other thread.) – Szabolcs May 31 '12 at 16:44
  • Ah, Greek version of Windows. Hmmm. – rm -rf May 31 '12 at 16:45