You can split values from a column and create a new column out of it in two ways -

 

1) By using suggestions on Right panel

 


Step 1: When you create a new connection, for example as shown here in Connect To hadoop article, you will get to Preview Screen.

 

Step 2: On preview screen, select the column in which you want to split something. As and when you highlight some values within the column, you can get a list of suggested split which is possible on that column in suggestions tab of right panel. 

  

Step 3 : Right Suggestion panel  will give you options to split on, from or between particular position, characters, separator etc. When you hover on these options, You will get a preview of the possible output adjacent to the column in light yellow color. 

 

Step 4: You can select the one which you like and click on it. It will apply the selected split operation on the column and will split the column in to two name "split" for the result . You can rename the column as desired.

 

 



2) By using drop down from Transformation section


Step 1: When you create a new connection, for example as shown here in Connect To hadoop article, you will get to Preview Screen.

 

Step 2: On preview screen, select the column in which you want to split and click on down arrow.

 

Step 3: You can either choose : "Split value by position" or "Split value by separator" by clicking on required option from String operations section. It will open up a pop up window for split.

 


 

Step 4: Fill the required details as listed below and click "Apply"

  • split by separator: The separator with which you want to split the column (if you select split by separator in drop down list)
  • split by position : The position number from which you want to split the column in to two (if you select split by position from the drop down list)

 


 


 

Step 5: Once you click on Apply, the system will apply the split operation on column considering the position or separator selected and column will be split in to two new column accordingly.