Hi Sean, thank you for your answer.

Some questions before I can give a suggestion:

> TEST_ID    JFI_ID    F_NAME    N1_NAME    N2_NAME    MODE    CHARACTERIST
> IC    PROPERTY
> 1    1    FLOW_1    NODE_A    NODE_B    MODE_1    HIGH FLOW    > 10 litres  /
> sec
> 1    1    FLOW_1    NODE_A    NODE_B    MODE_1    HIGH FLOW    > 50 gals /
> hour
> 1    1    FLOW_1    NODE_A    NODE_B    MODE_1    LOW LOSS    Pump current <
> 30
> 2    2    FLOW_2    NODE_C    NODE_D    MODE_1    LOW FLOW    < 5 litres / sec
> 44    3    FLOW_1    NODE_C    NODE_D    MODE_2    LOW FLOW    < 5 litres / sec

> Q1. How can I group and LIST() to produce a single row for the first test case
> (FLOW_1 from NODE_A to NODE_B with MODE_1). The result would look
> something like this:
> 1    1    FLOW_1    NODE_A    NODE_B    MODE_1    HIGH_FLOW, LOW_LOSS    >
> 10 litres / sec, > 50 gals / hour, Pump current < 30

Q: Did you mean to exclude the 2nd "HIGH_FLOW" characteristic value from the summary?

Yes, the same information should not be repeated within a single 'field'. The second "HIGH_FLOW" characteristic is only in the original result set because it has 2 properties (> 10 litres / sec and > 50 gals / hour - sorry contrived I know!) via the J_CHAR_PROPERTY table.

Q: Do you require the relationship/relative position between the "characteristic" and "property" values be maintained in the summary result?

It would be nice, but not essential. Currently the 'properties' of a 'characteristic' are not ordered to keep the example simple, but I could add a SEQ field to define the ordering and in that case it would be good for the LIST() to observe the order.

Mike