[peterlaursen]

here

http://dev.mysql.com/doc/mysql/en/charset-general.html

charset's and collations are explained.

[peterlaursen]

and BTW charset settings have a “top-down” approach.

Th server has a default. If not another default is specified for a DB, the server's default applies. If not another default is specified for a a table the default of the DB applies etc …

So I think PhPAdmin could have done it much better

[peterlaursen]

Quote:

I think the primary purpose of setting the encodings is to make sorting functions work properly

the primary purpose of COLLATIONS is to make sorting functions and such things work properly, I believe. The primary purpose of CHARSET is to assoicalte a graphic image a 'letter' or a 'sign' to a binary value. That is as of 4.1 ++. before 4.1 MySQL to keep both functionalities in CHARSET. But of course there is a cooperation os the OS-settings and the MySQL server on how thing s should be presented graphically ?

For instance German telephone books uses a very special sorting. MySQL created a german collation only for having the ability to sort the same way.

to my best knowing 🙂

BTW: If you put your japanese strings in a 'latin1' table and I read it from a Danish windows' version, how do they look then on my screen ?

[peterlaursen]

I voted YES!

I think that we are quite a few useers here that have been so much in contact, that we almost know one another. A place for a more unformal communication would be great

Also someone could come across some inforamtion about MySQL that does not directly relate to SQLyog, but nevertheless is of interest to SQLyog users. It could be an article in some media, a MySQL bug report …

BTW: I drink Thy Classic Økologisk (organic) preferably. Info about the organic beers from “Thisted Bryghus” here: http://www.thisted-bryghus.dk/katalog7.htm

It is perfect with a Highland Park single malt whisky from the Orkneys …

[peterlaursen]

still character and collation can be defined on a per table or per column basis

If for instance you want to create a japanese-danish dictonary application it would be relevant to have different setting for different colums.

I don't believe it a crazy example. International companies could have names of various items (that they sell for instance) in various languages.

But I have no idea of how to implement in SJA and the TUNNELing file.

[peterlaursen]

1.) If PK found, do a sync using PK.

2.) If PK is not available and UNIQUE INDEX is present, SJA will use the UNIQUE INDEX.

3.) If there are multiple UNIQUE INDEXES on a table, SJA will use the first UNIQUE INDEX.

change to

1.) If identical PK's found, do a sync using PK.

2.) If PK's are not available or not identical and identical UNIQUE INDEXes are present, SJA will use the UNIQUE INDEX.

3.) If there are multiple UNIQUE INDEXES on a table on both hosts, SJA will use the first UNIQUE INDEX identical on both hosts.

… what I also believe was what you meant!

[peterlaursen]

A question:

Can't CHARSET set be specific for each DB, table and even column ?

And what about COLLATION with MySQL 4.1 ++

[peterlaursen]

Thx 😀

Actually I think that the ODBC-import should have an option too, where users could enter the constraintname.

A possible way to do it would be to have a button “edit constraints” close to the FK checkbox (see attached) throwing up the SQL. From here you would be able to change the constraintname and even choose only SOME and not ALL of the FK's to be imported.

I know that support for constraintname must be implemented in SJA-code and the XML-schema too. How complicated that would be I don't know!

[peterlaursen]

Did you read this paper of mine?

http://www.webyog.com/forums/index.php?act=ST&f=2&t=1507

Quote:

I really only require it for this feature and I cannot alter the db tables.

Well, however ANY sync mechanism will need to be able to UNIQUELY identify each row of data. A Primary Key is the most simple way. In my paper I proposed an option to use any unique index as well. That should work too. But if you really can't alter the tables at all, that would not help you either. And besides I don't think that feature is “just around the corner”.

I think you are doing a mistake. Maybe YOU are able to identify which row of data on each host that correspond to one another. That is then because it is your data and you know them. However a computer will need a PK or a unique index to find out which row of data on each host that correspond.

What your are requesting is a magician, not a program 😛

[peterlaursen]

Well I understand Jason another waythen you then!

I'll let Linux users take over here!

Any Linux user here who CAN connect to localhost or 127.0.0.1 from the SJA ??

[peterlaursen]

Quote:

used 127.0.0.1 in place of the string localhost in the xml file

worked for other user. Would not that be OK ?

If it works for you too, it must be because of the way the OS resolves “localhost”.

Maybe using socket and not tcp ?

(I don't know much about Linux!)

[peterlaursen]

Quote:

you will not be able to execute it from SQL WINDOW too

well yes I think I was! But I'll look deeper into it.

It became all too confused yesterday night!

[peterlaursen]

Quote:

Are you referring to how the Notification Services Wizard executes the queries when clicking the Next button?

Yes that was what I meant. But I am not sure I understand everything perfectly yet!

Might spend some time working with a few more simplified test cases .. just for my own understanding of it!

[peterlaursen]

Yes you are right .. it is an issue with GUI functionality only.

From cmd.line/batch-mode there is no problem.

[peterlaursen]

hmmm …

One workaround could be creating new table with every query from one JOB.

then run another JOB to clean up.

hmmm … still thinking

read my MOTTO <_<

[Author]

Posts