A table with parts (numbers and names) and a table with amounts and an
action ID for receipts, sales and losses etc.

table parts:    ID , PARTNO, PARTNAME
table bookings  : BOOKID, ACTION, AMOUNT,  REPORTDATE

ACTION=1 for sales, ACTION=2 for receipts
ACTION=3 for losses, ACTION=4 for returns

How can I generate a list for a certain date that looks like this?

BOOKNO  , BOOKNAME , SALES, RECEIPTS, LOSSES, RETURNS
345678  , red Roses, 16   , 12      , 0     , 2
345679  
...

How can I sum up the same column with all different conditions in a
single SQL statement?

In the first statement, use multiple SELECT statements concatenated into a
single result set with a UNION join. Each SELECT retrieves a subset of the
total rows in the table, based on the value in your ACTION column. Each has
the same number of columns, but provides a value of 1 (one) for one of the
columns SALES, RECEIPTS, LOSSES, and RETURNS, and a value of zero for the
rest of these columns. For example:

  /* Query for ACTION value of 1 (Sales) */
  SELECT BOOKID, PARTNAME, 1 AS SALES, 0 AS RECEIPTS,
    0 AS LOSSES, 0 AS RETURNS
  FROM "Bookings.dbf"
    INNER JOIN "Parts.dbf"
      ON (BOOKID = PARTNO)
  WHERE (ACTION = 1)
  UNION ALL
  /* Query for ACTION value of 2 (Receipts) */
  SELECT BOOKID, PARTNAME, 0, 1, 0, 0
  FROM "Bookings.dbf"
    INNER JOIN "Parts.dbf"
      ON (BOOKID = PARTNO)
  WHERE (ACTION = 2)
  UNION ALL
  /* Query for ACTION value of 3 (Losses) */
  SELECT BOOKID, PARTNAME, 0, 0, 1, 0
  FROM "Bookings.dbf"
    INNER JOIN "Parts.dbf"
      ON (BOOKID = PARTNO)
  WHERE (ACTION = 3)
  UNION ALL
  /* Query for ACTION value of 4 (Returns) */
  SELECT BOOKID, PARTNAME, 0, 0, 0, 1
  FROM "Bookings.dbf"
    INNER JOIN "Parts.dbf"
      ON (BOOKID = PARTNO)
  WHERE (ACTION = 4)
  ORDER BY 1, 2

Then save this query to a text file (with a .SQL filename extension),
FIRST.SQL for example. This allows it to be used as if it were a table in
the second statement.

In the second statement (executed from a TQuery or a saved query file)
aggregates the data from the first statement, grouping on the book ID and
name. Apply the SUM function to each of the columns SALES, RECEIPTS,
LOSSES, and RETURNS returned by the first query and it will produce a count
of each for every book.

  SELECT BOOKID, PARTNAME, SUM(SALES), SUM(RECEIPTS),
    SUM(LOSSES), SUM(RETURNS)
  FROM "FIRST.SQL"
  GROUP BY BOOKID, PARTNAME