Create SUM of field values from 2 Tables but choose most recent date (from either table) - sql

Question

I have two tables with the same types of columns: host, frequency, and date.
The host is the primary key.
I want it so that I can combine a table from say March with April and sum up their frequencies.
Here is what I should expect in my final output.
Let X be some primary key value host.
If X only exists in one table, use that row in the final result.
If X exists in both tables. Sum up both row's freq and select the most recent date. So if we had to compare between 4/20/19 vs 4/2/19, then 4/20/19 should be the date selected.
Suppose I had the following tables:
Table: Report_4.1.19
host freq date
A 15 4/1/2019
C 30 4/1/2019
Table: Report_3.1.19
host freq date
A 10 3/1/2019
B 20 3/1/2019
My ideal output should be the following:
Table: Result
host sum(freq) date
A 25 4/1/2019
B 20 3/1/2019
C 30 4/1/2019
Here's what I tried so far:
SELECT host,sum(freq),date
from
(
select
host,
freq,
date
from Report_4.1.19
union
select
host,
freq,
date
from Report_3.1.19
)
group by host
While my code does appear to achieve the intended result, I'm not sure if I've properly accounted for the date selection. What can I do to modify my code (if needed)?

Answer 1

Step one: use a date format supported by sqlite date and time functions, as those can be meaningfully ordered and compared, unlike what you have. YYYY-MM-DD works.
Step two: use one table for all reports (you already track the date as a column to tell them apart!) to keep things simple and play to the strengths of relational databases. Make the PK host, date instead of just host.
Step three: it then becomes a simple query:
SELECT host, sum(freq), max(date) AS date
FROM reports
GROUP BY host
ORDER BY host;

Answer 2

Your query is almost correct. Just use the max of the date column from your union.
SELECT host,sum(freq),max(date)
from
(
select
host,
freq,
date
from Report_4.1.19
union
select
host,
freq,
date
from Report_3.1.19
)
group by host
Note: If you wanted daily sums, you would use date directly and then add date to your group by clause.