Hi,
this is an example table i am working with

Company| Account| Project|Description
--------------------------------
3311 |9999 |8181 | abc
1100 |9999 |8080 | bcd
2200 |9999 |8181 | abc
3300 |9999 |8282 | cde
4400 |1111 |8080 | bcd
5500 |1111 |8080 | bcd
6600 |1111 |8080 | bcd
7700 |1111 | |
8800 |1111 | |
1234 |2222 |8888 | def
2345 |2222 |8888 | def
3456 |2222 |8888 | def
4567 |3333 | |
6789 |3333 | |
7890 |3333 | |
The description field comes from another table by joining them together.
I am supposed to be able to get a result where if all the projects under the same account are the same, I will only display one line indicating the account number and the project. So since accounts 2222 and 3333 have the same projects under them(empty fields are counted as one type of project),
I should be able to get the result:

Company| Account| Project|Description
--------------------------------
|2222 |8888|def
|3333 | |
And I will leave the company field just blank


However, if the projects under the same account are different, I will have to display each and everyone of the account with the company name.

This should be the result.
Company| Account| Project|Description
--------------------------------
3311 |9999 |8181|abc
1100 |9999 |8080|bcd
2200 |9999 |8181|abc
3300 |9999 |8282|cde
4400 |1111 |8080|bcd
5500 |1111 |8080|bcd
6600 |1111 |8080|bcd
7700 |1111 | |
8800 |1111 | |

So if i combine both case toether I should be able to see the table as shown:
Company| Account| Project|Description
--------------------------------
3311 |9999 |8181|abc
1100 |9999 |8080|bcd
2200 |9999 |8181|abc
3300 |9999 |8282|cde
4400 |1111 |8080|bcd
5500 |1111 |8080|bcd
6600 |1111 |8080|bcd
7700 |1111 | |
8800 |1111 | |
|2222 |8888|def
|3333 | |

I used this query while I am working on my computer which is using mySQL ver 4.0.13:

SELECT distinct a.COMPANY,
a.ACCOUNT,
a.PROJECT,
projectMap.DESCRIPTION
FROM myTable b, myTable a left join projectMap on a.PROJECT=projectMap.PROJECT
WHERE a.ACCOUNT = b.ACCOUNT and a.PROJECT<>b.PROJECT
UNION

SELECT
' ', myTable.PROJECT,
myTable.ACCOUNT,
projectMap.DESCRIPTON,
FROM
myTable left join projectMap on
myTable.PROJECT=projectMap.PROJECT group by myTable.ACCOUNT having count(distinct myTable.PROJECT)=1

However when I brought it over to another computer which was using ver 4.0.12, the results I got was wrong.
I will get an extra line where account 1111 is also grouped together.
Company| Account| Project|Description
--------------------------------
3311 |9999 |8181|abc
1100 |9999 |8080|bcd
2200 |9999 |8181|abc
3300 |9999 |8282|cde
4400 |1111 |8080|bcd
5500 |1111 |8080|bcd
6600 |1111 |8080|bcd
7700 |1111 | |
8800 |1111 | |
|1111 |8080|bcd <-------extra line???
|2222 |8888|def
|3333 | |

I think it is some bug in ver 4.0.12 that causes this, however , I cannot upgrade the version right now as I am afraid it may affect the rest of the queries which have been developed with this version. Is there anyway to correct this query to make it work?

Thanks.