Match phrase query in Elasticsearch - Step-by-Step Execution

Concept Flow - Match phrase query

Receive query phrase

↓

Analyze phrase into tokens

↓

Search documents

↓

Check token order and proximity

↓

Return matching documents

The match phrase query takes a phrase, breaks it into words, then finds documents where these words appear in the same order and close together.

Execution Sample

Elasticsearch

{
  "query": {
    "match_phrase": {
      "message": "quick brown fox"
    }
  }
}

This query finds documents where the field 'message' contains the exact phrase 'quick brown fox' in order.

Execution Table

Step	Action	Input/Condition	Result/Output
1	Receive query	Phrase: 'quick brown fox'	Start processing
2	Analyze phrase	Split into tokens: ['quick', 'brown', 'fox']	Tokens ready for search
3	Search documents	Look for documents with tokens in order	Candidate documents identified
4	Check token order	Tokens must appear in sequence	Documents with 'quick' followed by 'brown' then 'fox' selected
5	Check proximity	Tokens must be adjacent (default slop=0)	Only exact phrase matches kept
6	Return results	Documents matching phrase	Final matching documents returned
7	End	No more steps	Query complete

💡 Query ends after returning documents that contain the exact phrase in order with no extra words between.

Variable Tracker

Variable	Start	After Step 2	After Step 4	Final
query_phrase	null	'quick brown fox'	'quick brown fox'	'quick brown fox'
tokens	[]	['quick', 'brown', 'fox']	['quick', 'brown', 'fox']	['quick', 'brown', 'fox']
candidate_docs	[]	[]	[docs with tokens in order]	[docs with exact phrase]

Key Moments - 2 Insights

Why does the query only return documents with the exact phrase and not documents with the words scattered?

Because the match phrase query checks that tokens appear in the exact order and adjacency (step 5 in execution_table), so scattered words do not match.

What happens if the phrase has extra spaces or punctuation?

The phrase is analyzed into tokens (step 2), so punctuation is removed and spaces separate tokens, ensuring the search matches the word sequence.

Visual Quiz - 3 Questions

Look at the execution table, what tokens are created after analyzing the phrase?

A ['quick brown fox']

B ['fox', 'brown', 'quick']

C ['quick', 'brown', 'fox']

D ['quick', 'fox']

💡 Hint

Check Step 2 in the execution_table where phrase is split into tokens.

At which step does the query ensure tokens appear in the correct order?

A Step 3

B Step 4

C Step 5

D Step 6

💡 Hint

Look at Step 4 in execution_table where token order is checked.

If the slop parameter was increased, how would the execution_table change?

A Step 5 would allow tokens to be farther apart

B Step 2 would create more tokens

C Step 4 would ignore token order

D Step 6 would return no documents

💡 Hint

Slop affects proximity check in Step 5.

Concept Snapshot

Match phrase query syntax:
{
  "query": {
    "match_phrase": { "field": "phrase" }
  }
}

Behavior: Finds documents where words appear in exact order and adjacency.
Key rule: Tokens must be in sequence with no extra words between (default slop=0).

Full Transcript

The match phrase query in Elasticsearch takes a phrase and searches documents where the phrase appears exactly as given. First, the phrase is split into tokens (words). Then Elasticsearch searches for documents containing these tokens in the same order and next to each other. This ensures only documents with the exact phrase match are returned. The process ends after returning these matching documents. Key points include tokenizing the phrase and checking token order and proximity. Increasing slop allows some words between tokens but by default, the phrase must be exact.

Practice

1. What does a match_phrase query in Elasticsearch do?

easy

A. Searches only for single words, ignoring phrases.

B. Searches for any of the words in the query, regardless of order.

C. Searches for an exact sequence of words in the given order.

D. Searches for documents with numeric values only.

Solution

1. Step 1: Understand the purpose of match_phrase

   The match_phrase query looks for the exact sequence of words in the order they appear in the query.

2. Step 2: Compare with other query types

   Unlike match query, which finds any word in any order, match_phrase requires the phrase to appear exactly.

3. Final Answer:

   Searches for an exact sequence of words in the given order. -> Option C

4. Quick Check:

   match_phrase = exact word order search [OK]

Hint: Remember: match_phrase means exact words in exact order [OK]

Common Mistakes:

• Confusing match_phrase with match query
• Thinking it ignores word order
• Assuming it works only for single words

2. Which of the following is the correct syntax for a match_phrase query in Elasticsearch?

easy

A. { "match": { "field": "search text" } }

B. { "match_phrase": { "field": "search text" } }

C. { "term": { "field": "search text" } }

D. { "range": { "field": { "gte": 10 } } }

Solution

1. Step 1: Identify the correct query type

   The match_phrase query uses the key "match_phrase" with the field and the phrase to search.

2. Step 2: Check the syntax format

   The correct syntax is a JSON object with "match_phrase" as the key and the field name mapping to the phrase string.

3. Final Answer:

   { "match_phrase": { "field": "search text" } } -> Option B

4. Quick Check:

   Correct JSON syntax for match_phrase = { "match_phrase": { "field": "search text" } } [OK]

Hint: Look for "match_phrase" key with field and phrase inside braces [OK]

Common Mistakes:

• Using "match" instead of "match_phrase"
• Confusing with "term" or "range" queries
• Incorrect JSON structure or missing braces

3. Given this Elasticsearch query:

{ "query": { "match_phrase": { "description": "quick brown fox" } } }

Which documents will match?

medium

A. Documents containing the words "fox brown quick" in any order.

B. Documents containing any of the words "quick", "brown", or "fox" in any order.

C. Documents containing the phrase "quick fox brown" exactly.

D. Documents containing the phrase "quick brown fox" exactly in that order.

Solution

1. Step 1: Analyze the match_phrase query

   The query searches for the exact phrase "quick brown fox" in the field "description".

2. Step 2: Understand phrase matching behavior

   Only documents with the words in the exact order "quick brown fox" will match, not any other order or partial words.

3. Final Answer:

   Documents containing the phrase "quick brown fox" exactly in that order. -> Option D

4. Quick Check:

   Exact phrase match = Documents containing the phrase "quick brown fox" exactly in that order. [OK]

Hint: match_phrase requires exact word order in the phrase [OK]

Common Mistakes:

• Assuming word order does not matter
• Thinking partial word matches count
• Confusing with match query behavior

4. You wrote this query but it returns no results:

{ "query": { "match_phrase": { "title": "fast runner" } } }

Your documents contain the phrase "runner fast". What is the issue?

medium

A. The phrase order is incorrect; match_phrase requires exact order.

B. The field name "title" is invalid.

C. match_phrase does not support multi-word phrases.

D. The query syntax is wrong; it should use "match" instead.

Solution

1. Step 1: Check phrase order in query vs documents

   The query searches for "fast runner" but documents have "runner fast" which is reversed.

2. Step 2: Understand match_phrase behavior

   match_phrase requires the exact word order, so reversed order causes no matches.

3. Final Answer:

   The phrase order is incorrect; match_phrase requires exact order. -> Option A

4. Quick Check:

   Exact order matters in match_phrase = The phrase order is incorrect; match_phrase requires exact order. [OK]

Hint: Check word order carefully when match_phrase returns no hits [OK]

Common Mistakes:

• Ignoring word order importance
• Assuming match_phrase works like match
• Changing field name unnecessarily

5. You want to find documents containing the phrase "data science" but also allow a gap of one word between "data" and "science". Which match_phrase query option should you use?

hard

A. { "match_phrase": { "field": "data science", "slop": 1 } }

B. { "match_phrase": { "field": "data science", "boost": 1 } }

C. { "match": { "field": "data science", "slop": 1 } }

D. { "term": { "field": "data science" } }

Solution

1. Step 1: Understand phrase slop in match_phrase

   The "slop" parameter allows words to appear with gaps or reordered within a limit.

2. Step 2: Identify correct syntax for slop

   Using "slop": 1 inside the match_phrase query allows one word gap between "data" and "science".

3. Final Answer:

   { "match_phrase": { "field": "data science", "slop": 1 } } -> Option A

4. Quick Check:

   Use slop to allow gaps in match_phrase = { "match_phrase": { "field": "data science", "slop": 1 } } [OK]

Hint: Add "slop" to match_phrase to allow word gaps [OK]

Common Mistakes:

• Using slop with match instead of match_phrase
• Confusing boost with slop
• Using term query for phrase matching