Question

I have run a query using Eclipse from a Sybase db. I need to eliminate duplicate entries but the results have mixed types – INT and TEXT. Sybase will not do distinct on TEXT fields. When I Save All results and paste that into Excel some of the TEXT field bleeds into the INT field columns – which makes Excel -Remove Duplicates tough to do.

I am thinking I might create an alias for my query, add a temp table, select the distinct INT column values from the alias and then query the alias again, this time including the TEXT values. Then when I export the data I save it into Word instead. It would look like this:

SELECT id, text
FROM tableA, TableB
WHERE (various joins here...)
AS stuff

CREATE TABLE #id_values
(alt_id CHAR(8) null)

INSERT INTO #id_values
(SELECT DISTINCT id
FROM stuff)

SELECT id, text
FROM stuff a
WHERE EXISTS (SELECT 1 FROM #id_values WHERE b.alt_id = a.id )

If there was a way to format the data better in Excel I would not have to do all this manipulation on the db side.I have tried different formats in the Excel import dialog..import as tab-delimited, space-delimited with the same end result.

Additional information: I converted the TEXT to VARCHAR but I now need a new column which has up to 5 entries per id sometimes. ID -> TYPE is 1-many? The distinct worked on the original list but now I need to figure out how to show all the new column values in one row with each id. The new column is CHAR(4).

Now my original select looks like this:

SELECT DISTINCT id, CONVERT(VARCHAR(8192), text), type_cd
FROM TableA, TableB
...etc

And I get multiple rows again for each type_cd attached to an id. I also realized I don’t think I need the ‘b.’ alias in front of *alt_id*.

Also, regardless of how I format the query (TEXT or VARCHAR), Excel continues to bleed the text into the id rows. Maybe this is not a sql problem but rather with Excel, or maybe Eclipse.

Answer 1

You are limited in how much data you can past into an Excel cell anyway, so convert your text to a varchar:

SELECT distinct id, cast(text as varchar(255)) as text
FROM tableA, TableB
WHERE (various joins here...)

I’m using 255, because that is the default on what Excel shows. You can have longer values in Excel cells, but this may be sufficient for your purposes. If not, just make the value bigger.

Also, as a comment, you should be using the proper syntax for joins, which uses the “on” clause (or “cross join” in place of a comma).