8 Transforming data
- Chris Welford
- Zoe Baldwin
- Anthony George
- Fiona Sargeant (Unlicensed)
By the end of this chapter you will be able to:
- Use the attribute function replaceAll()
- Use the attribute function toUpper()
- Use the attribute function substring()
- Use the attribute function if()
- Use the internal variable _out
Reformat post codes to remove all spaces and make upper case
In your Introduction to Modelling analysis model, you will now reformat post codes to remove all spaces, and ensure that all letters are upper case:
- In your model, add a new table:
- Hover over Customer Details with Region, and press Create New Table.
- Give the new table the Name:
Reformatted Customer Details.
- Drag all attributes from to the new table Customer Details with Region.
- Open the settings for the Reformatted Customer Details table.
- In the Attributes section, add a new attribute:
- Press Add New
- In the New Table Attribute settings, set:
- Name:
ReformattedPostCode
- Leave the settings for the attribute as default (Type is String, Length is 50)
- Expression:
- Name:
- In the Attributes section, add a new attribute:
toUpper(replaceAll(in.PostCode, " ", ""))
- Press Apply and Close on the New Table Attribute settings.
Above the attribute's Expression is a help link that will open a page containing details of all attribute functions available in PhixFlow. Find the attribute functions you used above – toUpper() and replaceAll() – click each of these to see the help pages specifically for these functions.
Inline help is also provided for functions, this will appear as you type. At the bottom of the popup help is a link directly to the page for the function in question.
Generate short code for Counties
Now you will generate a short code for counties by taking the first 3 letters of the county for each customer detail record:
- In the Reformatted Customer Details table settings, add a new attribute called
CountyShortCode.
- Leave the default settings.
Set the attribute's Expression to:
substring(in.County, 1, 3)
Press Apply and Close.
Categorise bills as "LARGE" or "NORMAL"
To categorise bills as "LARGE" or "NORMAL":
- In the Reformatted Customer Details table settings, add a new attribute called
BillCategory
- Leave the default settings.
Set the attribute's Expression to:
if (in.BILLAMOUNT > 10.0, "LARGE", "NORMAL")
Press Apply and Close.
Set special processing flag _out
Finally, you will now set a special processing flag. This will be set to "YES" when the last bill for the customer is large. To do this, we need to access an attribute value that is in the table, but not in any of the input pipes to the table, only the table output.
To do this, you will use the internal variable _out:
- In the Reformatted Customer Details table settings, add a new attribute called
SpecialProcessing
- Leave the default settings.
Set the attribute's Expression to:
if (_out.BillCategory == "LARGE", "YES", "NO")
Press Apply and Close.
Check Point - Table Attributes
- The Attributes section for the Reformatted Customer Details table should now look like this:
Run table
- In the settings for the table Reformatted Customer Details, press Apply and Close.
- Run analysis on the table Reformatted Customer Details.
- Have a look at the data in this table – check that all your transformations have been correctly applied, like the image below:
- Remember to save your model layout, Save.