Top 100 MongoDB Indexes Interview Questions with Crisp Answers

Basics of Indexes

1. What is an index in MongoDB?
   A data structure that improves query performance by allowing fast lookups instead of full collection scans (COLLSCAN).

2. What is the default index in MongoDB?
   Every collection has a unique index on the _id field that cannot be dropped.

3. How do indexes improve performance?
   They reduce the number of documents examined and disk I/O by storing sorted keys and pointers to documents.

4. What is the cost of indexes?
   Extra storage and write overhead — every insert/update/delete must update all relevant indexes.

5. How do you create an index?
   db.collection.createIndex({ field: 1 }) (1 = ascending, -1 = descending).

6. How do you list all indexes on a collection?
   db.collection.getIndexes() or db.collection.getIndexes().length for the count.

7. How do you drop an index?
   db.collection.dropIndex("indexName") or db.collection.dropIndex({ field: 1 }).

8. Can you drop the _id index?
   No. The _id index is mandatory and cannot be dropped.

9. What data structure does MongoDB use for indexes?
   B-tree (implemented by WiredTiger).

10. What is index selectivity?
    The ratio of distinct values to total documents. High selectivity (more unique values) makes indexes more effective.

11. What happens if no index exists for a query?
    MongoDB performs a full collection scan (COLLSCAN).

12. What is a covered query?
    A query that can be satisfied entirely from the index without fetching full documents (IXSCAN + no FETCH stage).

13. How do you check if a query uses an index?
    Use .explain("executionStats") and look for IXSCAN (good) vs COLLSCAN (bad).

14. What does totalDocsExamined tell you?
    Number of documents examined. Ideally close to nReturned for efficient queries.

15. What is the working set?
    The portion of data and indexes that must fit in RAM for optimal performance.

Index Types

16. What is a single-field index?
    An index on one field. Direction (1 or -1) does not matter for single-field indexes because MongoDB can traverse in either direction.

17. What is a compound index?
    An index on multiple fields. Field order is critical.

18. What is a multikey index?
    An index automatically created when indexing an array field. One index entry is created per array element.

19. What is a text index?
    Supports full-text search on string content using $text operator. One text index per collection.

20. What is a geospatial index?
    Supports location-based queries. 2dsphere for spherical (Earth) data, 2d for planar data.

21. What is a hashed index?
    Indexes the hash of a field’s value. Useful for equality queries and shard key distribution. Does not support range queries.

22. What is a wildcard index?
    Indexes all fields or all fields matching a pattern ($**). Useful for flexible/schemaless documents.

23. Can a compound index contain a hashed field?
    No. Hashed indexes cannot be part of a compound index.

24. How many fields can a compound index have?
    Up to 32 fields.

25. What is the maximum size of an index key?
    1024 bytes.

26. What is a unique index?
    Enforces uniqueness on the indexed field(s). Duplicate values cause E11000 duplicate key error.

27. What is a sparse index?
    Only indexes documents that contain the indexed field (skips missing/null fields). Legacy option; prefer partial indexes.

28. What is a partial index?
    Indexes only documents that match a filter expression (partialFilterExpression). More flexible than sparse.

29. What is a TTL index?
    Automatically expires documents after a specified time (expireAfterSeconds). Only works on date fields.

30. What is a hidden index?
    An index that is maintained but not used by the query planner. Useful for testing index removal safely.

31. What is a collation index?
    Supports language-specific string comparison rules (case-insensitive, accent-insensitive, etc.).

32. Can you create an index on an embedded document field?
    Yes, using dot notation: { "address.city": 1 }.

33. Can you index the entire embedded document?
    Yes, but queries must match the exact document structure (order and fields).

34. What is a case-insensitive index?
    Created using a collation with strength: 2 (or similar).

35. How many text indexes can a collection have?
    Only one text index per collection.

36. What operators does a text index support?
    Primarily $text and $meta for text score.

37. What is a 2dsphere index used for?
    Geospatial queries on spherical Earth data (GeoJSON points, lines, polygons).

38. What is a 2d index used for?
    Legacy geospatial index for planar (flat) coordinate data.

39. Can multikey indexes be compound?
    Yes, but a compound multikey index can contain at most one array field.

40. Do multikey indexes support covered queries?
    Only if the query projects only the indexed fields and the array field is not projected as an array.

Compound Indexes & Query Patterns

41. What is the index prefix rule?
    A compound index on {A:1, B:1, C:1} can be used for queries on A, A+B, or A+B+C. It cannot be used for B or C alone.

42. What is the ESR rule?
    Equality → Sort → Range. Optimal field order for compound indexes: equality fields first, then sort fields, then range fields.

43. Why does field order matter in compound indexes?
    MongoDB uses the index from left to right. After a range condition, remaining fields cannot be used efficiently for sorting.

44. Can a compound index support sorting?
    Yes, if the sort direction matches the index direction (or the reverse of the entire index).

45. What happens if sort direction does not match the index?
    MongoDB may perform an in-memory sort (SORT stage), which can be expensive or hit the 32 MB limit.

46. Can you have multiple compound indexes that share prefixes?
    Yes, but it is often redundant. Prefer one well-designed compound index.

47. How do you index for a query with equality and range?
    Put equality fields first, then the range field.

48. How do you index for a query that filters and sorts?
    Follow ESR: equality fields → sort field → range fields.

49. What is index intersection?
    MongoDB can use multiple indexes for a single query (usually less efficient than a single compound index).

50. When is index intersection useful?
    Rarely preferred. A well-designed compound index is almost always better.

51. Can compound indexes be used for queries on trailing fields?
    No. Only prefixes of the index can be used.

52. What is the difference between {a:1, b:1} and {b:1, a:1}?
    They support different query patterns. The first supports queries starting with a; the second supports queries starting with b.

53. How many compound indexes should you create?
    As few as possible. Each index adds write overhead. Design for your most common query patterns.

54. Can you use a compound index for a single-field query?
    Yes, if the single field is the prefix of the compound index.

55. What is the recommended way to name indexes?
    Use explicit names: createIndex(..., { name: "status_date_idx" }) for easier management.

Multikey, Text, Geospatial & Special Indexes

56. When is a multikey index created?
    Automatically when you index a field that contains an array in any document.

57. Can you create a multikey index on multiple array fields in one compound index?
    No. A compound multikey index can include at most one array field.

58. How do multikey indexes affect query performance?
    They can increase index size significantly if arrays are large.

59. What is the difference between indexing an array field and indexing array elements?
    Indexing the array field creates a multikey index with one entry per element.

60. Can text indexes be combined with other index types?
    Limited. Text indexes have restrictions when combined with other fields.

61. What language does a text index use by default?
    English. You can specify default_language or per-document language.

62. How do you create a text index on multiple fields?
    createIndex({ title: "text", content: "text" }).

63. What is the text score?
    A relevance score returned by $meta: "textScore" for ranking text search results.

64. Can you create a geospatial index on a non-GeoJSON field?
    Yes for 2d (legacy), but 2dsphere requires GeoJSON.

65. What is the difference between $near and $geoWithin?
    $near returns results sorted by distance; $geoWithin returns documents within a shape (no sorting).

66. What is a hashed index primarily used for?
    Even distribution of data across shards when used as a shard key. Also supports equality queries.

67. Does a hashed index support range queries?
    No.

68. What is a wildcard index useful for?
    Collections with dynamic or unpredictable field names (schemaless data).

69. How do you create a wildcard index on all fields?
    createIndex({ "$**": 1 }).

70. How do you create a wildcard index on specific fields?
    createIndex({ "attr.$**": 1 }).

71. Can wildcard indexes be sparse?
    Yes, they can use sparse: true or partial filter expressions.

72. What is the difference between sparse and partial indexes?
    Sparse skips missing fields. Partial indexes only documents matching a custom filter expression.

73. When should you use a partial index?
    When only a subset of documents needs to be indexed (e.g., only active users).

74. Can a partial index be unique?
    Yes. Uniqueness is enforced only on documents that match the partial filter.

75. How often does the TTL monitor run?
    Approximately every 60 seconds. Expiry is not exact.

Index Options & Constraints

76. What happens when you insert a duplicate into a unique index?
    The operation fails with error code 11000 (duplicate key error).

77. Can unique indexes contain null values?
    Yes, but only one document can have a missing/null value for a sparse unique index. Non-sparse unique indexes treat null as a value.

78. What is the difference between sparse: true and partialFilterExpression?
    Sparse is a special case of partial. Partial is more powerful and recommended.

79. Can you create a TTL index on a non-date field?
    No. TTL indexes require a date field (or an array of dates).

80. Can you have multiple TTL indexes on a collection?
    Yes, but each document can only be expired by one TTL index.

81. What is a background index build?
    Older versions built indexes in the background. Modern MongoDB builds indexes in a way that minimizes locking.

82. What is the difference between foreground and background index builds?
    Foreground builds lock the database. Background builds allow reads/writes but take longer. (Mostly historical in recent versions.)

83. How do you create a unique sparse index?
    createIndex({ email: 1 }, { unique: true, sparse: true }).

84. What is collation strength?
    Determines comparison sensitivity. Strength 1 = base characters only, Strength 2 = case-insensitive, Strength 3 = case-sensitive.

85. Can you change an existing index’s options?
    No. You must drop and recreate the index.

Query Optimization & Explain

86. What stages in explain() indicate a good index usage?
    IXSCAN followed by limited FETCH (or no FETCH for covered queries). Avoid COLLSCAN and large SORT.

87. What is a blocking sort?
    An in-memory sort that occurs when the index cannot provide the required order. Limited to 32 MB by default.

88. How do you force a particular index?
    Use .hint("indexName") or .hint({ field: 1 }).

89. What does executionStats.nReturned mean?
    Number of documents returned by the query.

90. What does executionStats.totalKeysExamined mean?
    Number of index keys examined.

91. What is a covered query advantage?
    No document fetch is needed, resulting in the fastest possible query performance.

92. How can you make a query covered?
    Project only fields that exist in the index (including _id if needed, or exclude _id).

93. What is the impact of $where on indexes?
    $where generally cannot use indexes efficiently and forces a collection scan.

94. Can regular expressions use indexes?
    Only prefix (anchored) regexes can use indexes. Unanchored regexes (/abc/) cannot.

95. What is the best way to monitor index usage?
    Use $indexStats aggregation stage or db.collection.aggregate([{ $indexStats: {} }]).

Best Practices & Advanced Topics

96. What is the biggest risk of having too many indexes?
    Write amplification — every write must update all indexes, increasing latency and storage.

97. Should you index every field you query?
    No. Index only fields used in high-frequency, selective queries. Prefer compound indexes.

98. What is the recommended order for compound index fields?
    Follow ESR: Equality fields first, Sort field next, Range fields last.

99. How do you safely remove an index in production?
    First hide the index (hideIndex), monitor performance, then drop it if safe.

100. What should you check before creating a new index?
     Query patterns (via profiler or logs), existing indexes, cardinality/selectivity, and write impact. Always validate with explain().