This [answer][1] motivated think again my following tables but I am still uncertain how to store statistic data for quick selection. 
I am designing [here][2] quick selections in big data where note that I am not finished how to store efficiently events of one measurement session into an index. 
I have three tables: *measurements*, *events* and *headers* where all have some static (never changes) columns about sizes in bytes:

- originalMeasurementSizeInBytes in *measurements* - static, size of original .raw -file 
- measurementHeaderSizeInBytes in *headers* - static, never changes

Example of my table schema *measurements* without names

    SERIAL                     | NOT NULL
    INTEGER                    | NOT NULL
    INTEGER                    | NOT NULL
    TIMESTAMP WITH TIME ZONE  | DEFAULT CURRENT_TIMESTAMP

where I am worried about increasing the schema size by INTEGER fields because of my need in statistics. 
However, I am not sure if it is better to create another table for measurement statistics.

How can static statistics stored efficiently in PostgeSQL? 

How can dynamic statistics stored efficiently in PostgeSQL? 
For instance, removal of one event would decrease the measurement size which includes events and header. 



  [1]: http://dba.stackexchange.com/a/23933/69807
  [2]: http://stackoverflow.com/q/31204011/54964