Kueri Term dan Match

Welcome to the fundamentals of querying data in Elasticsearch! Today, we'll explore two essential query types: term and match.

Understanding their differences is crucial for effective search. We'll learn when to use each to get exactly the results you need, whether it's an exact value or a flexible full-text search.

Before we query, let's index some simple documents into an index called products. We'll use these examples throughout the lesson.

Run these curl commands in your terminal (assuming Elasticsearch is running on localhost:9200):

The term query is used for finding exact values in a field. It looks for a precise match without any text analysis (like lowercasing or stemming).

• It's ideal for structured data like product IDs, categories, tags, or status fields.
• It works best with fields that are mapped as keyword (non-analyzed strings) or numeric types.

Let's find all products with the exact category "Electronics". Notice we're querying category.keyword because the default text field would be analyzed.

Try this curl command:

When you index a string field (like category), Elasticsearch often creates two versions by default:

• category: A text field, which is analyzed (broken into words, lowercased).
• category.keyword: A keyword field, which is treated as a single, exact string.

The term query expects an exact match, so using the .keyword sub-field ensures your query string isn't analyzed and matches the exact stored value.

The match query is your go-to for full-text search. Unlike term, it's designed to be smart about text.

• It performs text analysis on your search query.
• It breaks your query string into individual terms, lowercases them, and sometimes stems them (e.g., "running" -> "run").
• It then finds documents that contain these analyzed terms, and also assigns a relevancy score.

Let's search for products with "powerful laptop" in their description. Even if the document contains "A powerful and lightweight laptop", the match query will find it.

Try this curl command:

When you use a match query, Elasticsearch's text analysis process kicks in:

1. Tokenization: "powerful laptop" becomes "powerful" and "laptop".
2. Lowercasing: "Powerful" becomes "powerful".
3. Stop Words: Common words like "a", "the" might be removed (depending on the analyzer).

This makes match queries highly flexible for natural language search, finding relevant results even with minor variations in phrasing or case.

Here's a quick summary of when to use each query type:

• term query:
  - For exact value matching.
  - Does NOT perform text analysis on query string.
  - Best for keyword fields, IDs, precise filters.
• match query:
  - For full-text search.
  - PERFORMS text analysis on query string.
  - Best for text fields, search bars, descriptive content.

Great job! You've learned the fundamental difference between term and match queries.

• term is for surgical, exact value searches.
• match is for flexible, full-text searches leveraging powerful text analysis.

Knowing when to use each is key to building effective search experiences. Next, we'll explore how to combine these and other queries to create even more sophisticated search logic!