It appears that you want to report for distinct counts of the labels across both columns, in a single column. If so, then we can try the following union approach:

SELECT val, COUNT(*) AS cnt
FROM
(
    SELECT ColumnA FROM yourTable
    UNION ALL
    SELECT ColumnB FROM yourTable
) t
GROUP BY val;

WITH Items AS
(
    SELECT ColumnA As Item FROM MyTable
    UNION
    SELECT ColumnB FROM MyTable
)
SELECT Item
   , (SELECT COUNT(*) FROM MyTable WHERE ColumnA = Item) 
   , (SELECT COUNT(*) FROM MyTable WHERE ColumnB = Item) 
FROM Items

See it work here:

https://dbfiddle.uk/uHaP4yoq

One option is to use Union to get distinct labels and inner join labels to your table on either col_a or col_b and do the conditional aggregation using Case expressions:

--    S a m p l e    D a t a :
Create Table tbl (col_a varchar(4), col_b varchar(4));
Insert Into tbl (col_a, col_b)
VALUES row('KLAX', 'KONT'),
       row('KBUR', 'KJFK'),
       row('KSAN', 'KBUR'),
       row('KJFK', 'KPHX');

--      S Q L :
Select l.lbl, 
       Max(Case When l.lbl = t.col_a Then 1 Else 0 End) as cnt_a, 
       Max(Case When l.lbl = t.col_b Then 1 Else 0 End) as cnt_b
From  ( Select col_a as lbl From tbl   UNION   Select col_b From tbl ) l
Left Join tbl t ON( l.lbl In(t.col_a, t.col_b) )
Group By l.lbl

/*    R e s u l t :
lbl   cnt_a  cnt_b
----  -----  -----
KLAX      1      0
KBUR      1      1
KSAN      1      0
KJFK      1      1
KONT      0      1
KPHX      0      1  */

… and if you want to add total label count column ….

--      S Q L :
Select l.lbl, 
       Max(Case When l.lbl = t.col_a Then 1 Else 0 End) as cnt_a, 
       Max(Case When l.lbl = t.col_b Then 1 Else 0 End) as cnt_b,
       --
       Sum(Case When l.lbl = t.col_a Then 1 Else 0 End + 
           Case When l.lbl = t.col_b Then 1 Else 0 End) as cnt_lbl_total
From  ( Select col_a as lbl From tbl   UNION   Select col_b From tbl ) l
Inner Join tbl t ON( l.lbl In(t.col_a, t.col_b) )
Group By l.lbl

/*      R e s u l t :
lbl   cnt_a  cnt_b  cnt_lbl_total
----  -----  -----  -------------
KLAX      1      0              1
KBUR      1      1              2
KSAN      1      0              1
KJFK      1      1              2
KONT      0      1              1
KPHX      0      1              1  */

See the fiddle here.

Union the two data sets, then aggregate:

select code, max(cola) as a, max(colb) as b
from
(
  select columna as code, 1 as cola, 0 as colb from mytable
  union all
  select columnb as code, 0 as cola, 1 as colb from mytable
)
group by code
order by code;