Business Context
An online electronics retailer wants to implement an auto-complete feature on their homepage search bar. The goal is to suggest valid product categories and names (e.g., 'Wireless Headphones', 'Smart TV') as the user types to improve conversion rates and user experience.
About the Set : searchAnalytics
Data indexing and search functionalities.
Discover all actions of searchAnalytics Data Preparation
Creating a dataset of product descriptions and generating the initial term index required for the autocomplete build.
Copied!
| 1 | DATA mycas.products; LENGTH product_desc $100; INFILE DATALINES truncover; INPUT product_desc $ &; DATALINES; |
| 2 | Wireless Bluetooth Headphones Black |
| 3 | Noise Cancelling Over-Ear Headphones |
| 4 | Smart 4K Ultra HD LED TV |
| 5 | Portable Waterproof Bluetooth Speaker |
| 6 | Gaming Laptop 16GB RAM |
| 7 | Digital SLR Camera Kit |
| 8 | ; RUN; |
| 9 | |
| 10 | PROC CAS; search.buildTermIndex / TABLE={name='products'} docId='product_desc' casOut={name='product_terms', replace=true}; RUN; |
Étapes de réalisation
1
Execution of buildAutoComplete using the product term index.
Copied!
| 1 | |
| 2 | PROC CAS; |
| 3 | searchAnalytics.buildAutoComplete / index={name='product_terms'} casOut={name='product_autocomplete', replace=true}; |
| 4 | |
| 5 | RUN; |
| 6 |
2
Verification (Optional): Fetching a few rows to ensure the index structure is created.
Copied!
| 1 | |
| 2 | PROC CAS; |
| 3 | TABLE.fetch / TABLE={name='product_autocomplete'} to=5; |
| 4 | |
| 5 | RUN; |
| 6 |
Expected Result
The action successfully processes the 'product_terms' table and generates the 'product_autocomplete' table. This output table contains optimized n-grams or term fragments ready to be queried by a search interface for instant suggestions.