Data type Detection

Data type detection is used to detect the date type of feature (column). It verifies and finds Date-time, URL, Email, Currency, Phone, Latitude degree, latitude seconds, latitude minutes, longitude deg, longitude Seconds, and longitude minutes data types. It returns a list of dictionaries that contain the column name and its type.
Sample output:

    
copy
[{"name":"created_time","datatype":"DateTime"},{"name":"mail","datatype":"Email"},....]

DType Transformation

This ML operation generates new columns from URL, Email, Date-time and json type columns.

  1. URL Transformation

    This operation splits URL column into eight columns for better processing and performance.

    Sample input:

    link
    https://www.google.in/library?fetch=query#fragment_part

    Sample output:

    link_url_protocol https
    link_url_domain www.google.in
    link_url_path /library
    link_url_query fetch=query
    link_url_fragment fragment_part
    link_domain_tld in
    link_domain_country IN
    domain www
  2. Email Transformation

    This operation splits email column into three columns: first, middle, and last.
    Sample input:

    mail
    abc@zohocorp.com

    Sample output:

    mail_first mail_middle mail_last
    abc zohocorp com
  3. Date time Transformation

    This operation splits date time column into nine columns for better processing.
    Sample input:\

    dt
    2021-11-29 11:52:59

    Sample output:

    dt_day_of_week 1
    dt_date_no 29
    dt_year_no 2021
    dt_month_no 11
    dt_business_day 1
    dt_week_no_year 48
    dt_day_of_year 333
    dt_AM_PM AM
    dt_quarter_year 4
  4. Flatten Json

    This operation flattens the JSON value so that JSON contains only key-value pairs without nesting for better processing.
    Sample input:

    
copy
{ "user":{ "abc": {"id":1, "mail":"abc@zohocorp.com"} } }

Sample output:

    
copy
{ "user_abc_id":1, "user_abc_mail":"abc@zohocorp.com" }

Last Updated 2023-06-15 17:14:14 +0530 +0530