A set of values given: Assume table in SQL or list of dictionaries if using Python.
Basically a row of data contained information:
if it is post or it is a comment, row id and some other data.
Find distribution of comments.
#comments # posts
1 5000
2 6787
..
..

Question

A set of values given: Assume table in SQL or list of dictionaries if using Python. 
Basically a row of data contained information:
if it is post or it is a comment, row id and some other data.
Find distribution of comments. 
#comments # posts
1                       5000
2                       6787
..
..

Tide · Accepted Answer

Can also be with join, but I did it like this:

SELECT Count(a.parent_id), 
       num_com 
FROM   (SELECT id AS parent_id, 
               0  AS num_com 
        FROM   submission 
        WHERE  parent_id IS NULL 
               AND id NOT IN (SELECT parent_id 
                              FROM   submission 
                              WHERE  parent_id IS NOT NULL) 
        UNION 
        SELECT parent_id, 
               Count(id) AS num_com 
        FROM   submission 
        WHERE  parent_id IS NOT NULL 
        GROUP  BY parent_id) AS a 
GROUP  BY num_com;

TIde · Answer

For the above solution, assuming 1 table : submission {id, body, parent_id}

Sigiloso · Answer

I really think that the question is much simpler than this that Tide understood. Because if this problem wants the distribution of comments and posts and there is this information in the table (the guy said above), we just need to count...
I also believe that the distribution per day, or month, etc.

SELECT DAY(date) as per_day, content_type, count(content_type)
FROM t_post_comment
GROUP BY per_day, content_type

dc007 · Answer

From the way the question is worded here, it seems like these are the (relevant) variables in the table (table name - tableA)
1) type ('Comment', 'Post')
2) post_id

Seems like each row with 'Comment' entry will have a corresponding 'post_id' to identify which post it was a comment to.
Also each 'post' will have a corresponding 'post_id'
There may be user_id as well in the table, but let's keep things simple here.

Sql code to get #comments # posts
select comment_count, count(distinct post_id) as post_count
from 
(
select post_id, count(case when type = 'Comment' then 1 else 0 end) as comment_count
from tableA
group by post_id
)
group by comment_count

Meta

Pergunta de entrevista da empresa Meta

Respostas da entrevista

Empresas seguidas

Buscas de vagas