Processing of tables with WHEN.

[RyanFleck]

When two tables have an identical column, i.e. ID, and both tables are combined, the ID column from the second table is discarded without sorting.

I am certain that some gov't tables will have at least one shared column, and that data cannot be discarded.

ASSEMBLE overtime_table_A
  COLUMNS FROM table:hours
  COLUMNS FROM table:names;

Example rule here, see Section 2.

Current output is:

table:full_table
     0 | shared:              2 | english:            one | roman:             II |
     1 | shared:              3 | english:            one | roman:            III |
     2 | shared:              1 | english:            one | roman:              I |
     3 | shared:              2 | english:            two | roman:             II |
     4 | shared:              3 | english:            two | roman:            III |
     5 | shared:              1 | english:            two | roman:              I |
     6 | shared:              2 | english:          three | roman:             II |
     7 | shared:              3 | english:          three | roman:            III |
     8 | shared:              1 | english:          three | roman:              I |

Expected output is:

table:full_table
     0 | shared:              1 | english:            one | roman:              I |
     1 | shared:              2 | english:            two | roman:             II |
     2 | shared:              3 | english:          three | roman:            III |

[karfai]

In this example, the column with the same name is not used as a key for joining the tables and the data in the column is different in each table?

[RyanFleck]

No, the data in the column is identical, but stored in a different order. It could be used as a key to combine the data.

I assumed that, as any column could be matched, the rows would not always have the same data.

[karfai]

Ok... I'm not understanding what you're trying to do.... You want to retain both columns without using them as joining keys? If that's the case:

ASSEMBLE overtime_table_A
  COLUMNS FROM table:hours
  COLUMNS (employee_id AS names_employee_id, other, column) FROM table:names;

But... COLUMNS with an AS might not be working properly... If not, I'll add an issue.

[karfai]

In fact, table joining assemblies don't work well at all. Both of the steps below should produce the same table, but one produces nothing and the other a cross-product without filtering:

# No join at all
ASSEMBLE full_table
  COLUMNS FROM table:a
  COLUMNS FROM table:b WHEN shared == @shared;

# empty table
ASSEMBLE half_table
  COLUMNS FROM table:a
  COLUMN roman FROM table:b WHEN shared == @shared;