database - How to calculate size of the one record in hive?

Question

I need to understand how much memory my table will use. I know the arithmetic mean of records. I know what types of records have. But I can't understand how to calculate mean/max size of the record. Is this possible to calculate it if I get mean/max size of every column of record?

Answer 1

get mean/max size of every column

Also you can analyze table and fetch stats for column.

Analyze table:

ANALYZE TABLE tablename COMPUTE STATISTICS FOR COLUMNS;

Use DESCRIBE FORMATTED:

desc formatted  tablename  column_name

It will print these figures:

data_type, min, max, num_nulls, distinct_count, avg_col_len, max_col_len, num_trues, num_falses

Unfortunately, it works for single column at a time and you need to repeat describe command to get all columns stats.