How to overcome the encoding problem?

Question

linux

Essence of the question.
There is an ORacle DB, it is created with cl8iso8859p5 (so it was historically added). For a long time, all patch rolls were made from Windows machines, where the set NLS_LANG=AMERICAN_AMERICA.CL8MSWIN1251 variable was pre-registered in the console and everything rolled perfectly.
Now there was a need to roll a patch from a Linux machine.
When pulling script files from SVN, the files get ISO-8859-1 encoding, and Russian characters are obtained in the form of krakozyabr, in the same form they go to the database.
locale is set to UTF-8 by default.
empirically picked up an encoding in which everything is displayed correctly.
iconv -f cp1251 package_name.sql.
It is necessary to convert to UTF-8, I thought, and did so.
When I started to roll scripts, I ran into an insertion problem. Those values ​​that were previously correct suddenly ceased to fit into the table column. Somewhere I found that this is an effect due to UTF-8. Now I'm in vtupore, solving one problem, I get a bunch more.
Can anyone come across similar problems of translating encodings and scripts?

Answer 1

Install and try like this:

export NLS_LANG=AMERICAN_AMERICA.UTF8
export LC_ALL=ru_RU.utf8

Answer 2

this is the effect due to UTF-8

That's right, one UTF-8 character occupies two bytes. Therefore, the size of all text fields will have to be multiplied by two in this case.