Completion suggester in Elasticsearch - Step-by-Step Execution

Concept Flow - Completion suggester

User types prefix

↓

Elasticsearch receives prefix

↓

Completion suggester queries index

↓

Find matching suggestions

↓

Return suggestions to user

↓

User sees autocomplete options

The completion suggester takes a typed prefix, searches the index for matching suggestions, and returns them to the user for autocomplete.

Execution Sample

Elasticsearch

{
  "suggest": {
    "song-suggest": {
      "prefix": "nir",
      "completion": { "field": "suggest" }
    }
  }
}

This query asks Elasticsearch to suggest completions for the prefix 'nir' using the 'suggest' field.

Execution Table

Step	Action	Input/State	Output/Result
1	Receive prefix input	User types 'nir'	Prefix = 'nir'
2	Query completion suggester	Prefix = 'nir'	Search 'suggest' field for matches starting with 'nir'
3	Find matching suggestions	Index contains 'nirvana', 'nina simone', 'nirvikalpa'	Matches found: 'nirvana', 'nina simone', 'nirvikalpa'
4	Return suggestions	Matches found	Suggestions returned: ['nirvana', 'nina simone', 'nirvikalpa']
5	User sees suggestions	Suggestions returned	Autocomplete options shown to user

💡 Execution stops after suggestions are returned to the user.

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 4	Final
prefix	''	'nir'	'nir'	'nir'	'nir'	'nir'
matches	[]	[]	[]	['nirvana', 'nina simone', 'nirvikalpa']	['nirvana', 'nina simone', 'nirvikalpa']	['nirvana', 'nina simone', 'nirvikalpa']
suggestions	[]	[]	[]	[]	['nirvana', 'nina simone', 'nirvikalpa']	['nirvana', 'nina simone', 'nirvikalpa']

Key Moments - 3 Insights

Why does the completion suggester only return suggestions starting with the prefix?

Because the completion suggester is designed to match terms that start with the given prefix, as shown in execution_table step 3 where only matches starting with 'nir' are found.

What happens if no matches are found for the prefix?

If no matches are found, the matches list remains empty and no suggestions are returned, as would be seen if step 3 had an empty matches array.

Can the completion suggester return partial matches inside words?

No, it only returns suggestions that start with the prefix, not partial matches inside words, as demonstrated in step 3 where only terms starting with 'nir' are matched.

Visual Quiz - 3 Questions

Look at the execution_table, what is the value of 'matches' after step 3?

A []

B ['nirvana']

C ['nirvana', 'nina simone', 'nirvikalpa']

D ['simone', 'nirvikalpa']

💡 Hint

Check the 'matches' variable in the variable_tracker after step 3 and the output in execution_table row 3.

At which step does Elasticsearch return the suggestions to the user?

A Step 2

B Step 4

C Step 3

D Step 5

💡 Hint

Look at the 'Action' and 'Output' columns in execution_table row 4.

If the user typed 'ni' instead of 'nir', how would the matches change?

A More matches starting with 'ni' would appear

B Fewer matches would appear

C Matches would be the same

D No matches would appear

💡 Hint

Consider how prefix length affects matches in the completion suggester as shown in execution_table step 3.

Concept Snapshot

Completion suggester syntax:
{
  "suggest": {
    "name": {
      "prefix": "text",
      "completion": { "field": "fieldname" }
    }
  }
}

It returns autocomplete suggestions starting with the prefix.
Only terms starting with prefix are matched.
Useful for fast, prefix-based search suggestions.

Full Transcript

The completion suggester in Elasticsearch works by taking a prefix typed by the user, then searching the index for terms that start with that prefix in a special completion field. The process starts when the user types a prefix like 'nir'. Elasticsearch receives this prefix and queries the completion suggester on the specified field. It finds all matching suggestions that start with 'nir', such as 'nirvana', 'nina simone', and 'nirvikalpa'. These suggestions are then returned to the user to show autocomplete options. The execution table shows each step from receiving input to returning suggestions. Variables like 'prefix', 'matches', and 'suggestions' track the state through the process. Key points include that only terms starting with the prefix are matched, no partial matches inside words are returned, and if no matches exist, no suggestions are returned. The visual quiz tests understanding of the matches after searching, when suggestions are returned, and how changing the prefix affects matches. The concept snapshot summarizes the syntax and behavior of the completion suggester for quick reference.