全部来自表 A 并为表 B 中的每个集合包含一行，即使表 B 中没有相关记录

Question

I had a hard job describing this simple problem...

I want to return a series of data points for a stacked bar chart for each stack Label and Bar, whether or not there is a record for the Label/Bar datapoint, a null value being returned where there is no data record. The 'missing' Label must also be included:

    Label   Bar  Value
    a       X    99
    a       Y    23
    a       Z    97
    b       X    null
    b       Y    32
    b       Z    null

I have table DataLabel that contains the Labels for the each type of datapoint in the bar stack and a table DataSeries that contains the X-axis Labels and the Y-axis values to be calculated for each datapoint.

    create table DataLabel
    (id number
    , label varchar2(10));

    Insert into DataLabel values (1, 'a');
    Insert into DataLabel values (2, 'b');

    Create table DataSeries
    (id number
    , DataLabel_id number
    , val number
    , DataSeriesName varchar2(5));

    Insert into DataSeries values (1, 1, 99, 'X');
    Insert into DataSeries values (11, 1, 23, 'Y');
    Insert into DataSeries values (12, 2, 32, 'Y');
    Insert into DataSeries values (21, 1, 97, 'Z');

The problem arises, where the DataSeries table be does not contain a corresponding value for a DataLabel for a given datapoint.

I have managed an inelegant solution that does return the desired result in SQLfiddle using a Union join, but is there a better solution....?

Answer 1

Use a cross join to generate the rows and then a left join to bring in the values:

select dl.label, dsn.dataseriesname,
       ds.val
from datalabel dl cross join
     (select distinct dataseriesname from dataseries
     ) dsn left join
     dataseries ds
     on ds.datalabel_id = dl.id and
        ds.dataseriesname = dsn.dataseriesname
order by dl.label, dsn.dataseriesname;

Here is the SQL Fiddle.

Answer 2

A cross join among tables containing a correlated subquery might be an option:

select distinct label, DataSeriesName as bar,
      (select val 
         from DataSeries 
        where DataSeriesName=s.DataSeriesName 
          and DataLabel_id=l.id) as value
  from DataSeries s
 cross join DataLabel l 
 order by label, bar

Demo