Results and export

The CLI returns the candidate set with one extra column for every declared decision column.

+----+--------+-------+-------------+
| id | weight | value | is_selected |
+----+--------+-------+-------------+
| 1  | 5      | 44.61 | 1.0         |
| 2  | 20     | 22.13 | 0.0         |
| 3  | 12     | 98.80 | 1.0         |
+----+--------+-------+-------------+

optimal · objective 389.79 · highs

Candidate rows stay visible

Every candidate row is returned by the CLI, including rows the solver did not choose. Filter rows with a decision value of 1.0 in your next processing step.

Candidate columns retain their candidate-set order. Decision columns follow them.

Numeric decision values

Decision values are printed as floating-point numbers, including columns declared as INTEGER or 0/1 decisions. This is why a selected value appears as 1.0.

A decision column grouped with BY (...) repeats the group’s single selected value on every row in that group.

Solver report

The final line reports three values:

1. solver status, such as optimal
2. objective value
3. solver backend, such as highs

Infeasible and unbounded models return no result table. The process exits non-zero with an error.

Export CSV

Use --export to write the same CLI rows to a CSV file while keeping the printed result.

./deql \
  --register-files inventory.csv \
  candidates.deql \
  query.deql \
  --export solution.csv

The output contains one header row for the complete file.

solution.csv

id,weight,value,is_selected
1,5,44.61,1.0
2,20,22.13,0.0
3,12,98.8,1.0

Flight SQL returns a different shape

Through Flight SQL, 0/1 decision columns filter the result instead of being returned. Read the interface guide before switching a client from CLI output.

Arrow Flight SQL results Learn about filtered rows and solver metadata in Arrow.