[peterlaursen]

Quote:

Shouldn't the log be a little more descriptive of the error/issue/problem???

Everybody agrees on that. Too much work to do for too few people basically, I think!

Send a fresh XML-file and the log-file (with no 'fakes') to [email protected]!

And I also think, that the people at Webyog should now have detailed information about the source database (not just an 'ISAM-database' and the ODBC-driver! it it the metatable-structure of that DB that is interesting!).

But I'll let Ritesh take over!

[peterlaursen]

@Ritesh

didn't you forget a link? 😀

http://www.databasejournal.com/features/my…cle.php/3550146

[peterlaursen]

See attached!

This is what I get when pressing after highlighting a table in Object Browser.

This SQL

Code:

select * from `test2`.`table1` limit 0, 50

is executed to get this display. You can copy it from the HISTORY tab.

But you request that this SQL-statement is pasted into the SQL-pane ?

[peterlaursen]

At the top of the Forums (1st category)

http://www.webyog.com/forums/index.php?sho…st=0&#entry7439

[peterlaursen]

Quote:

I couldn't connect to MYSQL at first, though, after rebooting my PC, it goes well!

Probably the TCP-stack got messed up in Windows, or something like that. I don't think it is a SQLyog issue.

[peterlaursen]

Maybe this is a little bit easier to overview:

Two similar databases (a MySQL 4.1 and a 5.0 database) as shown in SQLyog Object Pane.

For the 4.1 database observe:

It the listing at the top the NUL- setting is not explicitly “NO” for column “ID”, The “create table” statement however has “NOT NUL in the definition”. And it is correct that MySQL adminstrator shows the column as NOT NUL (the adm.jpg is a shot of the 4.1 database)

For the 5.0 database the setting in the listing at the top IS explicitly “NO” for column “ID”.

…. hmmmm

[peterlaursen]

I think I may have found out something by now!

The “betcode”-column has NULL undefined in source and NO in target.

That seems to be one more difference (or a bug?) of MySQL version 4.1 and 5.0.

Null-definition is NO for autoincrement columns with 5.0 and undefined with 4.1. It is not the SQLyog copy too that causes this, I belive. It occurs like that everywhere with the two MySQL versions.l

I would not mind to work more into it, but let us build an “plan of attack” together with Ritesh …

[peterlaursen]

I am sorry … but you will have to guide me a little bit more!

To me it all looks identical.

But did you read this discussion:

http://www.webyog.com/forums/index.php?showtopic=1714

… and did you try the SQLyog 4.2 beta5 ?

Do you also have the NULL statements in the jobfile?

When you are sync'ing is then between identical MySQL-versions?

But I'd wait for Ritesh to reply too.

He knows exactly what they did between beta4 and beta5.

[peterlaursen]

I found this question in the FAQ sytem:

“I just bought sqlYog enterprise. 1) I copy a db from a 4.1.12 server to a 5.0RC server using sqlYog db/copy db.. 2) I compare using structure synch. Structure sync somehow does not seem to recognise columnd with NOT NULL set on the 5.0RC server. (they are however . I have checkd with Mysql ADmin). First I thought it was SQLYog that didnt copy it correct – but it seems to be something wring in the way SQLYog fetched this info from 5.0RC. Is this a bug or are I trying to do somethinsg that SQLYog are not ready to do yet? This also leads to that the database sync fails since the structures are recognised as different.”

I think this belongs in the Forums! I'll delete the entry from the FAQ – just for your information!

[peterlaursen]

@Shawn

Can you confirm that beta5 works with you?

[peterlaursen]

I cananot reproduce this. See attached picture.

Upper database is a MySQL 4.1.14.

Lower is a 5.0.13

Data have been copied from 4.1.14 to 5.0.13 and you see that everything is “NOT NULL” as it should be.

We will need create statements for the tables identical to yours. A a litte bit of samle data would be nice too …

[peterlaursen]

You should have it by now. And please note that there is no

Quote:

yes

in the file.

Shawn deleted all instances of

Quote:

NULL

And it worked. It did not before.

I still believe that we have two issues with MySQL verion 5 here!

1) an autoincrement have have a default

1) a default-value can't be NULL.

[peterlaursen]

I have it …

And you do in 5 minutes!

[peterlaursen]

Quote:

The problem arises when SQLyog specifies a DEFAULT value with auto_increment column and NULL as a default value.

Yup. Nobody were in doubt, I guess! But what we have been discussing her is WHY it happens. I have not been able to reproduce it with ACCESS as the source, so what is returned by the source DB when queried about table structure must be part of the problem?

Another part of the problem probably is that MySQL 5 is not so “relaxed” as previous versions in this respect. That is only one example. With this and future releases of MySQL we will se that MySQL approaches “Standard SQL” still more for each release. I for instance expect a much more 'strict' type-checking for the future. Will cause a lot of problems with clients and scripts that have been writtten for 3.23 and 4.0 …

@ Ritesh was it a bug with SQLyog in your opinion that has been fixed, or is it a workaround for diversities in SQL-dialects that have been implemented ?

And by the way, this

Quote:

The problem arises when SQLyog specifies a DEFAULT value with auto_increment column and NULL as a default value.

should be

Quote:

The problem arises when SQLyog specifies a DEFAULT value with auto_increment column or NULL as a default value.

and >> or !!!

The resource on the internet that Shawn found is about auto_increment columns. Shawn's own problem is 'NULL as a default value'. Not necessessarily 'NULL as a default value for auto_increment columns'.

[peterlaursen]

To get information about the table structure SQLyog performs a QUERY against the metadata tables of the source database. There is no other way. I think that the ODBC-source in this case returns a default = NULL where there really is NONE. Bad coding simply!

However, since this can cause problem with MySQL version 5, maybe SQLyog should implement som “filtering” for illegal defaults, such as NULL-defaults and defaults with an autoincrement variable. However, I don't have any idea of how widespread the problem is.

Let Ritesh have a look into the issue 🙄

[Author]

Posts