Select data from one table with the conditions from another

drop table if exists tablea;
create table tablea(id int,catid varchar(6),acca int,accb int,price int);

drop table if exists tableb;
create table tableb(catid varchar(6),name varchar(7),customer varchar(6), dt date);

truncate tablea; truncate tableb;

insert into tablea values
(2,'Order5',111,131,40),(3,'Order1',131,511,40),(4,'Order2',131,511,40),(5,'Order3',111,131,30),(6,'Order3',133,131,10);

insert into tableb values
(1,'Order1','Apple','2016-11-02'),(2,'Order2','Apple','2016-11-11'),(3,'Order3','Apple','2016-11-11'),(4,'Order4','Google','2016-11-11');

The first step in this solution is to create a dummy key (k) and decide if the row is a parent or child

MariaDB [sandbox]> select a.*,
    -> case when a.acca = 131 then 1
    -> else 2
    -> end as ParentOrChild,
    -> case when a.acca = 131 then concat(131 , year(b.dt), month(b.dt),day(b.dt),b.customer)
    -> else concat(a.accb, year(b.dt), month(b.dt),day(b.dt),b.customer)
    -> end as k
    -> from tablea a
    -> join tableb b on b.name = a.catid;
+------+--------+------+------+-------+---------------+------------------+
| id   | catid  | acca | accb | price | ParentOrChild | k                |
+------+--------+------+------+-------+---------------+------------------+
|    3 | Order1 |  131 |  511 |    40 |             1 | 1312016112Apple  |
|    4 | Order2 |  131 |  511 |    40 |             1 | 13120161111Apple |
|    5 | Order3 |  111 |  131 |    30 |             2 | 13120161111Apple |
|    6 | Order3 |  133 |  131 |    10 |             2 | 13120161111Apple |
+------+--------+------+------+-------+---------------+------------------+
4 rows in set (0.00 sec)

The next stage works out if the total child prices matches the parent price

MariaDB [sandbox]> select s.parentorchild, s.k,
    -> sum(case when s.parentorchild = 1 then s.price else 0 end ) -
    -> sum(case when s.parentorchild = 2 then s.price else 0 end ) MatchedPrice
    -> from
    -> (
    -> select a.*,
    -> case when a.acca = 131 then 1
    -> else 2
    -> end as ParentOrChild,
    -> case when a.acca = 131 then concat(131 , year(b.dt), month(b.dt),day(b.dt),b.customer)
    -> else concat(a.accb, year(b.dt), month(b.dt),day(b.dt),b.customer)
    -> end as k
    -> from tablea a
    -> join tableb b on b.name = a.catid
    -> ) s
    -> group by  s.k
    -> order by  s.k,s.parentorchild;
+---------------+------------------+--------------+
| ParentOrChild | k                | MatchedPrice |
+---------------+------------------+--------------+
|             1 | 13120161111Apple |            0 |
|             1 | 1312016112Apple  |           40 |
+---------------+------------------+--------------+
2 rows in set (0.00 sec)

We now know the the dummy key (k) that we are interested in (MatchedPrice = 0) so if we join back on dummy key we get the rows we are interested in

ariaDB [sandbox]> select u.id,u.catid,u.customer,u.dt,u.acca,u.accb,u.price
    -> from
    -> (
    -> select s.parentorchild, s.k,
    -> sum(case when s.parentorchild = 1 then s.price else 0 end ) -
    -> sum(case when s.parentorchild = 2 then s.price else 0 end ) MatchedPrice
    -> from
    -> (
    -> select a.*,
    -> case when a.acca = 131 then 1
    -> else 2
    -> end as ParentOrChild,
    -> case when a.acca = 131 then concat(131 , year(b.dt), month(b.dt),day(b.dt),b.customer)
    -> else concat(a.accb, year(b.dt), month(b.dt),day(b.dt),b.customer)
    -> end as k
    -> from tablea a
    -> join tableb b on b.name = a.catid
    -> ) s
    -> group by  s.k
    -> order by  s.k,s.parentorchild
    -> ) t
    -> join
    -> (select a.*, b.customer,b.dt,
    -> case when a.acca = 131 then 1
    -> else 2
    -> end as ParentOrChild,
    -> case when a.acca = 131 then concat(131 , year(b.dt), month(b.dt),day(b.dt),b.customer)
    -> else concat(a.accb, year(b.dt), month(b.dt),day(b.dt),b.customer)
    -> end as k
    -> from tablea a
    -> join tableb b on b.name = a.catid
    -> ) u
    -> on u.k = t.k
    -> where MatchedPrice = 0
    -> ;
+------+--------+----------+------------+------+------+-------+
| id   | catid  | customer | dt         | acca | accb | price |
+------+--------+----------+------------+------+------+-------+
|    4 | Order2 | Apple    | 2016-11-11 |  131 |  511 |    40 |
|    5 | Order3 | Apple    | 2016-11-11 |  111 |  131 |    30 |
|    6 | Order3 | Apple    | 2016-11-11 |  133 |  131 |    10 |
+------+--------+----------+------------+------+------+-------+
3 rows in set (0.00 sec)

Note it might be better for performance purposes if you used work tables instead of trying to do this in a single query.