Well, the other simple way to merge into a table is to do a
catalog/directory merge where you set up a single table row in your mail
merge main document, with one merge field in each cell.

But you haven't said how you are deciding where your output document is
going to be split, because in that case, you just get one continuous table
with no page breaks etc.

The reason I suggested a DATABASE field is because when you want continuous
tabular output of arbitrary length, that's the way to do it. If you need to
do a merge, then consider using one data source for the merge, with one
record for each "document" you want to produce, and use one DATABASE field
in the mail merge main document that extracts the records related to the
current record in the data source, typically using a key field nested in the
SELECT query in the DATABASE field.

E.g. your data source has records like

sk   stext
1    mytexta
2    mytextb
2    mytextc
3    mytextd
3    mytexte
3    mytextf

etc. and you want documents with

mytexta

then
mytextb
mytextc

then
mytextd
mytexte
mytextf

In that case, you can use a data source for the merge such as SELECT
DISTINCT sk FROM mytable, then put a DATABASE field in your Mail Merge Main
Document that does

SELECT stext FROM mytable WHERE sk = { MERGEFIELD sk }

(where { MERGEFIELD sk } is a merge field code and the {} are the special
field braces you can insert using ctrl-F9)

There are problems with this approach, e.g. in recent versions of Word MS
code gets confused between the fields in the data source for the merge and
the fields in tables in the DATABASE field, but you probably either need to
be doing that, or "rolling your own merge" by inserting table rows as
required using C#, in which case your current design won't work either.

Peter Jamieson