Like here, I need to dynamically add columns depended on the counts of occurrences.
The differences are:
- There is no orders column. No other table is joined or any foreign keys.
- Additional columns are fixed and user specified. In the following example, it is set to 2.
Example Table:
+--+-----+
|id|data | // + Any other Constant rows
+--+-----+
|1 |P | // 1. of id1
+--+-----+
|1 |Q | // 2. of id1
+--+-----+
|2 |R | // 1. of id2
+--+-----+
|3 |S | // 1. of id3
+--+-----+
|3 |a | // 2. of id3
+--+-----+
|3 |C | // 3. of id3
+--+-----+
Example result table with 2 data columns (which is pre set by user - of course it also could be 3 or more)
+--+-----+-----+
|id|data1|data2| // + all other constant columns
+--+-----+-----+
|1 |P |Q | // exactly 2 entries
+--+-----+-----+
|2 |R |NULL | // no second entry
+--+-----+-----+
|3 |S |a | // no third entry
+--+-----+-----+
My first idea was to create an identifying column, like the "orders" columns, but I couldn't find a way to accomplish that.
Is it possible to realize that? I prepared a sqlFiddle
