Business Context
A marketing agency tracks trending topics on social media. User-generated content is 'dirty', containing hashtags, special characters, mixed case, and excessive whitespace. The test ensures the action processes these irregularities without failing.
About the Set : searchAnalytics
Data indexing and search functionalities.
Discover all actions of searchAnalytics Data Preparation
Creating a dataset with mixed case, special characters (#, @), and null values.
Copied!
| 1 | DATA mycas.social_posts; LENGTH post_content $100; INFILE DATALINES truncover; INPUT post_content $ &; DATALINES; |
| 2 | #SummerVibes #2025 |
| 3 | !!! BREAKING NEWS !!! |
| 4 | @user_handle check this out |
| 5 | leading spaces test |
| 6 | |
| 7 | NOTE: The line above was empty |
| 8 | ; RUN; |
| 9 | |
| 10 | PROC CAS; search.buildTermIndex / TABLE={name='social_posts'} docId='post_content' casOut={name='social_terms', replace=true}; RUN; |
Étapes de réalisation
1
Attempting to build auto-complete index from irregular terms.
Copied!
| 1 | |
| 2 | PROC CAS; |
| 3 | searchAnalytics.buildAutoComplete / index={name='social_terms'} casOut={name='social_autocomplete', replace=true}; |
| 4 | |
| 5 | RUN; |
| 6 |
2
Saving the resulting index to a specific caslib (Testing casOut flexibility with specific path/lib).
Copied!
| 1 | |
| 2 | PROC CAS; |
| 3 | TABLE.promote / name='social_autocomplete' targetLib='casuser'; |
| 4 | |
| 5 | RUN; |
| 6 |
Expected Result
The action completes without error. Special characters are either indexed (if supported by the underlying engine) or stripped, but the process does not crash. The output table 'social_autocomplete' is created, demonstrating the action's robustness against dirty input data.