SUBSTITUTE Function

Contents:

Input can be specified as a column reference, a function returning a string, or a string literal, although string literal usage is rare.

• A column reference can refer to a column of String type.
• If no match is found, the function returns the source string.
• If multiple matches are found in a single string, all replacements are made.

Wrangle vs. SQL: This function is part of Wrangle , a proprietary data transformation language. Wrangle is not SQL. For more information, see Wrangle Language.

Basic Usage

Column reference example:

substitute(myURL,`{ip-address}`,myDomain)

Output: Searches the myURL column values for sub-strings that match valid IP addresses. Where matches are found, they are replaced with the corresponding value in the myDomain column. 

Function reference example:

substitute(upper(companyName),'ACME','New ACME')

Output: Searches the uppercase version of values from the companyName column for the string literal ACME. When found, these matches are replaced by New ACME in the companyName column.

Syntax and Arguments

substitute(string_source,string_pattern,string_replacement[,ignore_case, pattern_before, pattern_after])

For more information on syntax standards, see Language Documentation Syntax Notes.

string_source

Name of the item to be searched. Valid values can be:

• String literals must be quoted ( 'Hello, World' ).
• Column reference to any type that can be inferred as a string, which encompasses all values
• Functions that return string values

Multiple values and wildcards are not supported.

Usage Notes:

string_pattern

String to find. This value can be a string literal, a Trifacta® pattern, a regular expression, a column, or a function returning a String value.

• String literals must be quoted ('Hello, World').
• Multiple values and wildcards are not supported.

Usage Notes:

string_replacement

Value with which to replacement any matched patterns. Value can be a string, a function returning string values, or a column reference containing strings.

• String literals must be quoted ('Hello, World').
• Column reference to any type that can be inferred as a string, which encompasses all values. 
• Multiple values and wildcards are not supported.

Usage Notes:

ignore_case

When true, matches are case-insensitive. Default is false. 

Usage Notes:

pattern_before

String literal or pattern to find in a position before the pattern to match. 

Usage Notes:

pattern_after

String literal or pattern to find in a position after the pattern to match. 

Usage Notes:

Examples

Example - Partial obfuscation of credit card numbers

Source:

Suppose you have the following transactional data, which contains customer credit card numbers. 

For security purposes, you wish to redact the first three sets of digits, so only the last set of digits appears. 

Transformation:

To make the substitution, you must first change the type of the column to be a string:

You can then use the following transformation to perform the pattern-based replacement of four-digit sets that end in a dash with XXXX:

To indicate that the column no longer contains valid information, you might choose to rename it like in the following:

Results: