Chinese Characters cannot be dispalyed in data grid after upgraded to HeidiSQL 12

Hi all,

After I have upgraded to version 12, Chinese characters returned in data grid are shown as '?'. No such problem in version 11. Anythings can be done in settings to fix?

Thanks in advance.

Regards, Patrcik

That could be a font issue. Your both screens look as if you were using different fonts. Try the same font, in Tools > Preferences > Grid formatting.

Hi Ansgar,

Thanks for your response. However I checked the setting in Tools > Preferences > Grid formatting between two versions (11 and 12) and found no differences. They got records from the same server/database/table. Actually I did not alter anything after upgraded to version 12. Collate of the table is 'utf8_general_ci'. Are there any chances that somethings were hard-coded in version 12?

Regards, Patrick

That could be a font issue. Your both screens look as if you were using different fonts. Try the same font, in Tools > Preferences > Grid formatting.

Ok, so this is not a font issue.

There are two things which might cause that:

• since v12.0.0.6492, HeidiSQL does not change the connection's charset if it's already starting with "utf". But I did that after the 12.0 release, in the nightly builds. Are you on the release or on the latest nightly build?
• issue #1351 got anothe fix for critical characters in the grids (especially emojis and multibyte characters). But that did not affect the text values themselves, just the length of strings was counted wrong.

You should please check the connection charset, once you have a connection:

SHOW VARIABLES LIKE 'character_set%';

Hi Ansgar,

My Version 12 was already on the latest nightly build 12.0.0.6507, the problem persists. The results returned from "SHOW VARIABLES LIKE 'character_set%';" are different however:

Version 12.0.0.6507 (Displays Chinese characters as '?'):

Version 11.3.0.6321 (Displays Chinese characters correctly)::

But both HeidiSQL V11 and V12 clients connect to the same server/database.

Ok, I thought you were using the last 12.0 release. In the nightly builds, you are probably running into a new issue due to my change for issue #287.

Please restart your connection and watch out for such a line in the log panel at the bottom, saying "Changing character set from x to y":

Please post a screenshot or the text you see there.

Hi Ansgar，

No such line "/ Changing character set from latin1 to utf8mb4 /" as you mentioned. The screen shot is as below:

Strange. Your "character_set_client" variable is "latin1", but mysql_character_set_name() returns "utf8_mb4".

I recommend to downgrade to v12.0 until I find some fix for that.

Okay, I will fallback to earlier version. Am looking forward to your fix.

Many Thanks.

Strange. Your "character_set_client" variable is "latin1", but mysql_character_set_name() returns "utf8_mb4".

I recommend to downgrade to v12.0 until I find some fix for that.

Hi, I've the same problem with latin2 -> utf8mb4 in version 11, the characters are displayed correctly and in version 12 they are not.

After executing the instruction (SHOW VARIABLES LIKE 'character_set%';), we have the differences as shown in the attached picture.

Thanks for the comparison and screenshot. That's quite strange but I will find a fix.

I recommend either downgrading to v11.3, or fire some "SET NAMES utf8mb4;" after connecting.

libmysql.dll seems to cause this issue. At least I could reproduce it various times here. Changing the library in the session to "libmariadb.dll" fixed it for me. Please try the same and check if that helps.

Next build forces utf8mb4 charset when libmysql is used.

I installed the latest version 12.0.0.6512, I changed the library to libmysql.dll and the connected characters are displayed correctly. In the case of libmariadb.dll that I had set by default, they still display badly.

Thank you for your help.

Yes, using libmariadb.dll causes the problem, changing to libmysql.dll gets rid of the problem.

I installed the latest version 12.0.0.6512, I changed the library to libmysql.dll and the connected characters are displayed correctly. In the case of libmariadb.dll that I had set by default, they still display badly.

Thank you for your help.

I thought it was the other way around, that libmariadb.dll is using the right charset from the beginning. This gets really strange.

@patrickwong and @rokiko: which server versions are you running?

And can one please post the character-set related settings from your my.cnf/my.ini on the server?

I could finally reproduce it here, on a super old MySQL 5.1 server. I saw that in the very first screenshot in this thread. Why the heck do you use such old server versions? These have bugs, probably serious ones.

However, I will try to find out which server version was the first to report the correct charset through mysql_character_set_name, and then make an extra call for such old versions.

Hello!

Sorry to revive this post, but have you found a fix for it?

I'm also having problems with some of my DB's, which are still on MySQL 5.1 (I know...), because data on utf8 tables is shown incorrectly.

In my case they are not chinese characters, but regular leters with diacritics, e.g. accents and ñ. For instance, the name Sofía Peña Argüello shows wrong:

I'm already on the latest Heidi build, i.e. 12.1.0.6557, but the bug persists.

Now, as other comment suggested, I tried changing the connection library, from libmariadb.dll, to libmysql.dll, and it works. Now it shows them correctly:

But I don't know if there's a downside of using that library, or I may get bugs elsewhere (e.g. exporting data, etc). Also using an older version of Heidi (< 12.x) is not ideal, so a proper fix would be better.

Let me know if you need more info.

Thanks in advance!

Go and upgrade your old MySQL installation. MySQL 5.1 is much older than the libmysql.dll you are now using. There won't be any further workarounds and quirks in HeidiSQL for such old systems.