transformations:match
Table of Contents
MATCH
Category: Transform / Advanced
Description
This action goes through text values in the source table and finds matching values in a second "lookup" table (lookup table). Matched lookup values then replace original values.
Use cases
- The Fuzzy mode of this action can be used to match text values that may have typos.
- This can be used for keyword detection in customer support tickets.
- This action can be used for flagging fraud detection phrases in financial institutions.
- Content moderation/spam detection in social media platforms.
- Categorize products for an e-commerce site by converting long names to keywords within the names.
- Log file analysis for IT & Security by flagging keywords in system and network logs.
- Flag competitive brand names within customer comments.
Action settings
Setting | Description |
---|---|
Column | Select the column in the current dataset that contains the values to match in the lookup table. |
Lookup table | Select the dataset to be use for the lookup. |
Lookup column | This setting appears once the Lookup table has been selected. Select the column that contains the values to be matched to. |
Match mode | Options: Contains, Starts with, Ends with, and Fuzzy. See the table below. |
Case sensitive match | When checked, text case must match in the lookup values. |
If not matched | Choose how unmatched values are handled. Options: Keep existing value, Make empty, or Default to (and choose whether this value is an expression, or text or a number, and enter the expression or text/number value). |
Match mode settings
Mode | Description |
---|---|
Contains | A lookup value is returned if the source value contains it. For example, "tor" would be returned if the source value was "Motorola". |
Starts with | A lookup value is returned if the source value starts with it. For example, "Moto" would be returned if the source value was "Motorola". |
Ends with | A lookup value is returned if the source value ends with it. For example, "rola" would be returned if the source value was "Motorola". |
Fuzzy | A lookup value is returned if the source value is within the specified edit distance. The Damerau–Levenshtein edit distance measures the number of edit operations required to produce one word from another. |
Max. edit distance | This setting appears when the Match mode is set to "Fuzzy". Choose the edit distance used by the Fuzzy-matching process. See the Fuzzy mode, above. |
Remarks
In the case of multiple matches, an arbitrary match is taken (in no particular order) and other matches are ignored.
Examples
Example #1
Flag and categorize potential issues within customer comments.
Before (source table)
Table 1: Table of customer comments.
Comment |
---|
Item arrived broken. |
No comments. |
My order is late again! |
Table 2: Lookup table
Issue list |
---|
refund |
late |
broken |
After (result table)
Comment |
---|
broken |
No comments. |
late |
Action parameters
Column: Source text (Table 1)
Lookup table: Table 2
Lookup column: Issue list (Table 2)
Match mode: Contains
If not matched: Keep existing value
Example #2
Flag and categorize potential fraudulent transactions in a financial institution.
Before (source table)
Table 1: Text value to match.
Transaction summary |
---|
Wire transfer to unknown receipient |
Mobile check deposit |
Scheduled transfer from checking |
Table 2: Lookup table
Fraud flags |
---|
lottery |
wire |
gift card |
After (result table)
Transaction summary |
---|
wire |
Mobile check deposit |
Scheduled transfer from checking |
Action parameters
Column: Source text (Table 1)
Lookup table: Table 2
Lookup column: Fraud flags (Table 2)
Match mode: Contains
If not matched: Keep existing value
Example #3
Return a Lookup value contained within the Source text value.
Before (source table)
Table 1: Text value to match.
Source text |
---|
Motorola |
Table 2: Lookup table
Lookup values |
---|
tor |
Moto |
rola |
Motrola |
After (result table)
Source text |
---|
tor |
Action parameters:
Column: Source text (Table 1)
Lookup table: Table 2
Lookup column: Lookup values (Table 2)
Match mode: Contains
If not matched: Make empty
Community examples
- How to look up and replace a string of text (Project; Module: Main; Group: Tab 1; Table: Table 3; Action position: 3)
- Substring Query (Project; Module: Main; Group: Tab 1; Table: Table 1; Action position: 3)
- Iterate through a set of column (Project; Module: Main; Group: Tab 1; Table: in.csv; Action position: 5)
See also
transformations/match.txt · Last modified: 2025/02/10 18:41 by craigt