find "smart" command

[jqnatividad]

As opposed to the search command, the find command will be a "smart" command, using the stats cache along with a frequency cache (see #2075) to find values within the CSV that are based on the valid domain values for a column.

For example:
qsv find data.csv fruits_column apple

list all rows from data.csv where fruits_column=apple.

It will return no rows if fruits_column does not have the value apple.

So far, nothing new... but if we issue the command

qsv find data.csv fruits_column _top

It will return all rows where fruits_column is the "top" occuring value (i.e. the mode).

qsv find data.csv fruits_column _bottom returns all rows where fruits_column is equal to the least occurring value.

Further:
qsv find data.csv fruits_column _top3
returns all rows from the top 3 domain values for the fruits column (e.g. apple, strawberry, banana - the top three occurring values)

qsv find data.csv fruits_column _TOP3
on the other hand, only returns rows where the fruits column is equal to the third most occuring fruits_column value (e.g. banana)

The opposite behavior will behave in a symmetric manner for _bottom, _BOTTOM.

This was inspired by #2056

[jqnatividad]

If a regular expression is used, it will only be applied to the domain values of the column, as compiled in the frequency table cache.