Indexing and querying semi-structured documents using a key-value store

1. A computer-implemented method for processing queries based on documents, the method comprising:

storing a plurality of indexes based on documents, each document having a document identifier and comprising fields, wherein each field has field data comprising one or more of: a field name, a field value, and a field type, wherein each index comprises key-value pairs stored in a key-value store, each index associated with a respective set of attributes selected from: (1) a document identifier attribute, (2) a field name attribute, (3) a field value attribute, and (4) a field type attribute, the set of attributes combined for obtaining keys of key-value pairs of the index, the plurality of indexes comprising:

a document index configured to identify documents matching a criteria based on fields;

a field index configured to access fields of a given document; and

a columnar index configured to determine aggregate values associated with fields;

receiving, from a query source, a query requesting information stored in documents indexed using the plurality of indexes, the query specifying a search criteria;

selecting an index from the plurality of indexes for processing the query based on a type of the query;

generating a key-expression by combining the set of attributes corresponding to the selected index;

accessing the requested information using the selected index, the accessing comprising identifying key-value pairs matching the generated key-expression; and

transmitting the accessed information to the query source.

2. The computer-implemented method of claim 1 , further comprising:

receiving a document comprising one or more fields; and

updating the plurality of indexes based on the received document, the updating comprising, storing in each index, one or more key-value pairs based on fields of the received document.

3. The computer-implemented method of claim 1 , wherein updating the plurality of indexes comprises storing a key-value pair in the document index, the key-value pair comprising:

a key obtained by combining the field name, the field type, the field value, and a document identifier identifying the document.

4. The computer-implemented method of claim 1 , wherein the query requests documents having a specified value of a specified field type for a specified field name, wherein the key-expression is obtained by combining the specified field name, the specified field type, and the specified field value, wherein accessing the requested information comprises:

identifying key-value pairs matching the key-expression using the document index; and

identifying one or more documents corresponding to the matching key-value pairs.

5. The computer-implemented method of claim 4 , wherein the plurality of indexes store fields of a first document and a second document, wherein the first document includes a field having the specified field name with the specified field type, and the second document includes a field having the specified field name and a different field type, wherein the one or more identified documents corresponding to the matching key-value pairs include the first document but exclude the second document.

6. The computer-implemented method of claim 1 , wherein updating the plurality of indexes comprises, storing a key-value pair in the field index, the key-value pair comprising:

a key obtained by combining the document identifier and the field name, and

a value representing the field value.

7. The computer-implemented method of claim 1 , wherein the query requests one or more fields of a document and the search criteria specifies a document, wherein the key-expression comprises a document identifier for the specified document, wherein accessing the requested information comprises:

identifying key-value pairs matching the key-expression using the field index; and

accessing one or more fields of the document from the matching key-value pairs.

8. The computer-implemented method of claim 1 , wherein updating the plurality of indexes comprises, storing a key-value pair in the columnar index, the key-value pair comprising,

a key obtained by combining the field name, the field type, and the document identifier, and

a value representing the field value.

9. The computer-implemented method of claim 1 , wherein the query requests an aggregate of a field of a specified field name having a specified field type across a set of documents, wherein the key-expression is obtained by combining the specified field name and the specified field type, wherein accessing the requested information comprises:

identifying key-value pairs matching the key-expression using the columnar index; and

aggregating the values of the identified key-value pairs.

10. The computer-implemented method of claim 1 , further comprising:

receiving an updated document corresponding to the received document;

identifying one or more fields of the updated document that were modified compared to the received document; and

for each of the one or more identified fields, updating the key-value pairs of the key-value store corresponding to the identified field.

11. A non-transitory computer readable storage medium storing instructions that when executed by one or more computer processors, cause the one or more computer processors to performs steps comprising:

storing a plurality of indexes based on documents, each document having a document identifier and comprising fields, wherein each field has field data comprising one or more of: a field name, a field value, and a field type, wherein each index comprises key-value pairs stored in a key-value store, each index associated with a respective set of attributes selected from: (1) a document identifier attribute, (2) a field name attribute, (3) a field value attribute, and (4) a field type attribute, the set of attributes combined for obtaining keys of key-value pairs of the index, the plurality of indexes comprising:

a document index configured to identify documents matching a criteria based on fields;

a field index configured to access fields of a given document; and

a columnar index configured to determine aggregate values associated with fields;

receiving, from a query source, a query requesting information stored in documents indexed using the plurality of indexes, the query specifying a search criteria;

selecting an index from the plurality of indexes for processing the query based on a type of the query;

generating a key-expression by combining the set of attributes corresponding to the selected index;

accessing the requested information using the selected index, the accessing comprising identifying key-value pairs matching the generated key-expression; and

transmitting the accessed information to the query source.

12. The non-transitory computer readable storage medium of claim 11 , wherein the query requests documents having a specified value of a specified field type for a specified field name, wherein the key-expression is obtained by combining the specified field name, the specified field type, and the specified field value, wherein instructions for accessing the requested information comprise instructions for:

identifying key-value pairs matching the key-expression using the document index; and

identifying one or more documents corresponding to the matching key-value pairs.

13. The non-transitory computer readable storage medium of claim 11 , wherein updating the plurality of indexes comprises, storing a key-value pair in the field index, the key-value pair comprising:

a key obtained by combining the document identifier and the field name, and

a value representing the field value.

14. The non-transitory computer readable storage medium of claim 11 , wherein the query requests one or more fields of a document and the search criteria specifies a document, wherein the key-expression comprises a document identifier for the specified document, wherein instructions for accessing the requested information comprise instructions for:

identifying key-value pairs matching the key-expression using the field index; and

accessing one or more fields of the document from the matching key-value pairs.

15. The non-transitory computer readable storage medium of claim 11 , wherein updating the plurality of indexes comprises, storing a key-value pair in the columnar index, the key-value pair comprising,

a key obtained by combining the field name, the field type, and the document identifier, and

a value representing the field value.

16. The non-transitory computer readable storage medium of claim 11 , wherein the query requests an aggregate of a field of a specified field name having a specified field type across a set of documents, wherein the key-expression is obtained by combining the specified field name and the specified field type, wherein instructions for accessing the requested information comprise instructions for:

identifying key-value pairs matching the key-expression using the columnar index; and

aggregating the values of the identified key-value pairs.

17. The non-transitory computer readable storage medium of claim 11 , wherein the stored instructions are further for:

receiving an updated document corresponding to the received document;

identifying one or more fields of the updated document that were modified compared to the received document; and

for each of the one or more identified fields, updating the key-value pairs of the key-value store corresponding to the identified field.

18. A computer system comprising:

one or more computer processors; and

a non-transitory computer readable storage medium storing instructions that when executed by the one or more computer processors, cause the one or more computer processors to perform steps comprising:

storing a plurality of indexes based on documents, each document having a document identifier and comprising fields, wherein each field has field data comprising one or more of: a field name, a field value, and a field type, wherein each index comprises key-value pairs stored in a key-value store, each index associated with a respective set of attributes selected from: (1) a document identifier attribute, (2) a field name attribute, (3) a field value attribute, and (4) a field type attribute, the set of attributes combined for obtaining keys of key-value pairs of the index, the plurality of indexes comprising:

a document index configured to identify documents matching a criteria based on fields;

a field index configured to access fields of a given document; and

a columnar index configured to determine aggregate values associated with fields;

receiving, from a query source, a query requesting information stored in documents indexed using the plurality of indexes, the query specifying a search criteria;

selecting an index from the plurality of indexes for processing the query based on a type of the query;

generating a key-expression by combining the set of attributes corresponding to the selected index;

accessing the requested information using the selected index, the accessing comprising identifying key-value pairs matching the generated key-expression; and

transmitting the accessed information to the query source.

19. The computer system of claim 18 , wherein the query requests documents having a specified value of a specified field type for a specified field name, wherein the key-expression is obtained by combining the specified field name, the specified field type, and the specified field value, wherein instructions for accessing the requested information comprise instructions for:

identifying key-value pairs matching the key-expression using the document index; and

identifying one or more documents corresponding to the matching key-value pairs.

20. The computer system of claim 18 , wherein the query requests one or more fields of a document and the search criteria specifies a document, wherein the key-expression comprises a document identifier for the specified document, wherein instructions for accessing the requested information comprise instructions for:

identifying key-value pairs matching the key-expression using the field index; and

accessing one or more fields of the document from the matching key-value pairs.