N1QL New Features in couchbase 7.0

Confidential and Proprietary. Do not distribute without Couchbase consent. © Couchbase 2020. All rights reserved.
N1QL New Features
New features in the upcoming release
October, 2020
Keshav Murthy | VP, Couchbase R&D

Confidential and Proprietary.
Do not distribute without Couchbase consent. © Couchbase 2020. All rights reserved.
Agenda
01/
02/
03/
04/
05/
06/
N1QL So far… (short version)
Collections for N1QL
Language Features
Optimizer Features
Indexing Features
07/
Search Features
Transactions
08/ Wrap

Confidential and Proprietary. Do not distribute without Couchbase consent. © Couchbase 2020. All rights reserved. 3
N1QL So Far...
1

Give developers and
enterprises an expressive,
powerful, and complete
language for querying,
transforming, and
manipulating JSON data.
N1QL
SELECT breweries.name AS brewery,
COUNT(*) AS bwcount
FROM beers INNER JOIN breweries
ON beer.brewery_id = breweries.id
GROUP BY breweries.name;
INSERT INTO beers VALUES("paleale:12",
{"name":"Irish pale ale", "":"", "abv": 0.05});
UPDATE beers SET abv = 0.08
WHERE name = "Irish pale ale";
DELETE FROM breweries WHERE country = "Denmark";

N1QL = Productivity

Data
Service
Data
Service
Query
Service
Index
Service
FullText
Search
Couchbase Cluster
SDK
(KV, FTS,
Query)
SDK
(KV, FTS,
Query)
Web
Console
Analytics
Service
Eventing
Service
N1QL

7
SQL Features
Triggers
Stored Procedures in PL-SQL,
T-SQL, languages
JSON, XML
SQL is English
for Relational
Database
SQL Invented by Don
Chamberlin &
Raymond Boyce at
IBM
N1QL, based on
SQL, is English for
JSON
N1QL was invented by
Gerald Sangudi at
Couchbase
SQL
Instance
Database
Schema
Table
Row
Column
N1QL
Cluster
Bucket
Scope
Collection
Document
Attribute
SQL
Input and Output: Set(s) of
Tuples
N1QL STMT
CREATE BUCKET
CREATE INDEX
None
SELECT
INSERT
UPDATE
DELETE
MERGE
Subqueries
JOIN
GROUP BY
ORDER BY
OFFSET, LIMIT
EXPLAIN
PREPARE
EXECUTE
GRANT ROLE
REVOKE ROLE
INFER
PREPARE
EXECUTE
ADVISE (6.6)
ADVISOR
UPDATE-STATISTICS
CREATE FUNCTION
BEGIN WORK
COMMIT WORK
SAVEPOINT
ROLLBACK
EXECUTE
FUNCTION
Set of
JSON
N1QL Model
Set of
JSON
SQL
Tooling
ODBC, JDBC, .NET
Hibernate
BI Tools
erwin
TOAD
SQL Indexes
Primary Key
Secondary Key
Composite
Range Partitioned
Expression
Functional
Spatial
Search
SQL Logic
3 valued logic
TRUE, FALSE, NULL/UNKNOWN
N1QL
Tooling
ODBC,JDBC, .NET
Couchbase SDKs
Spring
BI Tools
erwin
SQL
TRANSACTIONS
ACID
Multi-Statement
Multi-table
Mullti-Database
Savepoints
Commit
Rollback [savepoint]
2 phase locking
Multi Isolation Levels
Single node
N1QL
TRANSACTIONS
ACID
Multi-Statement
Mullti-collection
Multi-Bucket
Savepoints
Commit
OCC (Optimistic)
READ COMMITTED
Distributed tx
SQL Datatypes
Numeric
Decimal
Character
Date Time
Timezone
BLOB
Spatial
JSON
N1QL
Datatype
Numeric
Boolean
Character
Array
Object
Null
JSON
Conversion Functions
SQL
Optimizer
Rule Based
Cost Based
Index Selection
Join reordering
Query Rewrites
NL, Hash, Merge join
N1QL
Optimizer
Rule Based
Cost Based
Index Selection
Join as specified
Query Rewrites(Ltd)
NL, Hash join
N1QL
Engine
(SCALE UP &
SCALE OUT)
SQL STMT
CREATE TABLE
CREATE INDEX
ALTER TABLE
SELECT
INSERT
UPDATE
DELETE
MERGE
Subqueries
JOIN
GROUP BY
ORDER BY
OFFSET, LIMIT
EXPLAIN PLAN
PREPARE
EXECUTE
GRANT
REVOKE
DESCRIBE
PREPARE
EXECUTE
TRUNCATE
N1QL
Input and Output:
Set(s) of JSON
Tuples
SQL Model
Set of
Tuples
SQL Engine
(SCALE UP)
N1QL Features
Eventing Service
UDFs (N1QL, Javascript)
Defacto JSON
N1QL Logic
4 valued logic
TRUE, FALSE,
NULL/UNKNOWN, MISSING
N1QL
Indexes
Primary
Secondary
Composite
Partial
Expression
Functional
Array Index
Search
Flex Index(6.6)

8
SQL Features
Triggers
T-SQL, languages
JSON, XML
SQL is English
for Relational
Database
SQL Invented by Don
Chamberlin &
Raymond Boyce at
IBM
N1QL, based on
SQL, is English for
JSON
Gerald Sangudi at
Couchbase
SQL
Instance
Database
Schema
Table
Row
Column
N1QL
Cluster
Bucket
Scope
Collection
Document
Attribute
SQL
Tuples
N1QL STMT
CREATE BUCKET
CREATE INDEX
None
SELECT
INSERT
UPDATE
DELETE
MERGE
Subqueries
JOIN
GROUP BY
ORDER BY
OFFSET, LIMIT
EXPLAIN
PREPARE
EXECUTE
GRANT ROLE
REVOKE ROLE
INFER
PREPARE
EXECUTE
ADVISE (6.6)
ADVISOR
UPDATE-STATISTIC
S
CREATE FUNCTION
BEGIN WORK
COMMIT WORK
SAVEPOINT
ROLLBACK
EXECUTE FUNCTION
Set of
JSON
N1QL Model
Set of
JSON
SQL Tooling
ODBC, JDBC, .NET
Hibernate
++
SQL Indexes
Primary Key
Secondary Key
Composite
Range Partitioned
Expression
Functional
Spatial
Search
SQL Logic
3 valued logic
N1QL Tooling
ODBC,JDBC, .NET
Couchbase SDKs
Spring
SQL
TRANSACTIONS
ACID
Multi-Statement
Multi-table
Mullti-Database
Savepoints
Commit
2 phase locking
Single node
N1QL
TRANSACTIONS
ACID
Multi-Statement
Mullti-collection
Multi-Bucket
Savepoints
Commit
OCC (Optimistic)
READ COMMITTED
Distributed tx
SQL Datatypes
Numeric
Decimal
Character
Date Time
Timezone
BLOB
Spatial
JSON
N1QL
Datatype
Numeric
Boolean
Character
Array
Object
Null
JSON
SQL
Optimizer
Rule Based
Cost Based
Index Selection
Join reordering
Query Rewrites
N1QL
Optimizer
Rule Based
Cost Based
Index Selection
Join as specified
Query Rewrites(Ltd)
NL, Hash join
N1QL
Engine
(SCALE UP &
SCALE OUT)
SQL STMT
CREATE TABLE
CREATE INDEX
ALTER TABLE
SELECT
INSERT
UPDATE
DELETE
MERGE
Subqueries
JOIN
GROUP BY
ORDER BY
OFFSET, LIMIT
EXPLAIN PLAN
PREPARE
EXECUTE
GRANT
REVOKE
DESCRIBE
PREPARE
EXECUTE
TRUNCATE
N1QL
Input and Output:
Set(s) of JSON
Tuples
SQL Model
Set of
Tuples
SQL Engine
(SCALE UP)
N1QL Features
Eventing Service
Defacto JSON
N1QL Logic
4 valued logic
TRUE, FALSE,
N1QL
Indexes
Primary
Secondary
Composite
Partial
Expression
Functional
Array Index
Search
Flex Index(6.6)

N1QL - New Features in the next release
● INDEXES
○ Indexes per collections
○ All the existing Index features on collections
○ Supports up to 10,000 GSI indexes in a cluster
○ Improved performance for indexes on collections
● FULL TEXT SEARCH
○ Search indexing & querying on collections
○ Index on a single collection or multiple collections
○ Flex index to power N1QL (6.6)
● TRANSACTIONS
○ Multi-document,scope,collection,bucket
○ Multi-Statement
○ BEGIN TRANSACTION
○ COMMIT
○ SAVEPOINT
○ ROLLBACK
● Data Model Abstractions
○ SCOPE
○ COLLECTIONS
● LANGUAGE
○ ADVISE, ADVISOR()
○ User Defined Functions
■ N1QL
■ Javascript
● OPTIMIZER
○ Cost Based Optimizer
○ UPDATE STATISTICS
○ Index Advisor
■ ADVISE
■ ADVISOR()
■ Enhanced for Cost Based Optimizer

Collections for N1QL
2

N1QL: Data Logical Containment
UDF foo()
UDF foo2()
_default
Couchbase Cluster (e.g.
Bucket (cxprof)
Scope (usa)
Collection
(profille)
Collection
(loginfo)
● Bucket contains one or more logical scopes
● Each scope contains collections & UDFs.
● Cluster->Bucket->Scope->Collection
● Cluster->Bucket->Scope->UDFs
● In N1QL, use the collection as a keyspace reference
● Use query_context to for shorter reference
● Fully Backward Compatible
○ All your bucket operations still work
○ Uses _default scope & collection
Scope (uk)
Scope (eu)
Scope
(_default)
SELECT l.profname AS name,
l.uspszip AS postalcode
FROM cxprof.usa.loginfo l where l.type = "login";
CREATE INDEX i1 ON
cxprof.usa.loginfo(profname, uspszip)
PARTITION BY HASH(META().id)

N1QL
● Extend all N1QL DML SELECT, INSERT, UPDATE,
UPSERT, DELETE, MERGE to work with Couchbase
Collections
● New DDLs for creating and dropping
scopes, collections and UDFs
● Support fully qualified path
bucket.scope.collection or relative path
with query_context
● RBAC GRANT, REVOKE support at bucket,
scope and collection level
● Indexes can be created on specific
collections. Index names have to be unique
per collection
● Supports up to 1,000 collections and 10,000
indexes in a cluster
CREATE SCOPE cxprof.uk;
CREATE COLLECTION cxprof.uk.users;
CREATE COLLECTION cxprof.uk.royals;
CREATE INDEX i1 ON cxprof.uk.users(name, postal);
GRANT SELECT ON cx.uk.users TO manchester;
CREATE INDEX i1 ON cxprof.uk.royals(name,county);
GRANT SELECT ON cx.uk.royals TO london;
CREATE FUNCTION default:cx.uk.getusers(p)
{( SELECT name
FROM cx.uk.users
WHERE postal = p)}
SELECT u.name
FROM default:cxprof.uk.getusers("NW8 8Q") AS u;

Language Features
3

N1QL: UDFs in N1QL Language
● New Statements
○ CREATE FUNCTION
○ EXECUTE FUNCTION
○ DROP FUNCTION
● Use it in ANY DML query: SELECT, INSERT,
UPDATE, UPSERT, DELETE, MERGE
● Allows zero or more parameters and return values.
● Returns can be single value or an array of values.
● Each function is still executed on the same node as
the statement
● All the functions are to be created under a scope
○ UDFs and Collections are at the same level,
hierarchically speaking
● Unsupported for CREATE INDEX
○ CREATE INDEX i1 on f(bk.sc.foo(a))
CREATE FUNCTION default:cx.uk.getusers(p)
{( SELECT name
FROM cxprof.uk.users
WHERE postal = p)};
SELECT u.name
FROM default:cxprof.uk.getusers("NW8 8Q") AS u;
CREATE FUNCTION default:cxprof.uk.volume(l, w, b)
{(l * w * b)}
SELECT default:cxprof.uk.volume(9, 23, 43);
INSERT INTO cxprof.uk.users(KEY UUID(), VALUE v)
SELECT default:cxprof.uk.volume(48, 4, 23)}) AS v
SELECT * FROM cxprof.uk.shapes
WHERE default:cxprof.uk.volume(x, y, z) > 1000;
EXECUTE FUNCTION default:cxprof.uk.volume(3, 2,
4);

N1QL User Defined Functions using Javascript
● Add the javascript into the library
● Add the javascript function into the lilbray.
● CREATE FUNCTION in your scope
● This function can be used in any DML
● Larger functions can be registered from a file.
curl -v -X POST
http://localhost:8093/functions/v1/libraries/math/
functions/add -H 'content-type: application/json'
-d '{"name": "add", "code": "function add(a, b) {
let data = a + b; return data; }"}' -u
Administrator:password
CREATE FUNCTION default:cxprof.uk.jsadd(x, y)
LANGUAGE JAVASCRIPT AS "add" at "math";
EXECUTE FUNCTION default:cxprof.uk.jsadd(99,232);
/* Error : Mismatch number of parameters*/
EXECUTE FUNCTION default:cxprof.uk.jsadd(1, 2, 3);
[
{
"code": 10104,
"msg": "Incorrect number of arguments supplied
to function jsadd - cause: jsadd"
}
]
● Each query node comes with V8 javascript engine
● Each invocation of the javascript has about 8ms
overhead.
○ That’s each invocation overhead.
● Do something substantial in the UDF.
● The number of parameters during CREATE
FUNCTION and invocation should match
● Function name should be UNIQUE within the scope
● Create up to 500 functions in a cluster.

N1QL User Defined Functions using Javascript
● You can have the code in a file and register
● Everything else is the same.
● Globals aren’t supported
● The javascript functions are for compute
○ They can’t invoke N1QL statements from within
● They can invoke the usual Javascript library
File: add3.js
{"name": "add3", "code": "function add3(a, b, c) {
let res = a + b + c; return res; }"}
curl -v -X POST
http://localhost:8093/functions/v1/libraries/math/
functions/add3 -H 'content-type: application/json'
-d @./add3.js -u Administrator:password
CREATE FUNCTION default:cxprof.uk.jsadd3(p,q,r)
LANGUAGE JAVASCRIPT AS "add" at "math";
SELECT default:cxprof.uk.jsadd3(3, 4, 5)

Optimizer Features
4

N1QL: Cost Based Optimizer (CBO)
32 minutes
RULE BASED ROUTING COST BASED ROUTING

N1QL: Rule Based Optimizer (RBO)
N1QL Parser Semantic Analyzer
Rule Based
Optimizer
Query Execution
Tree
Executor
Query Results
Debug info (Query
Proﬁle)
MetadataSystem metadata
Query Explain PlanN1QL Statement

N1QL: Cost Based Optimizer (CBO)
N1QL Parser Semantic Analyzer
Cost Based
Optimizer
Query Execution
Tree
Executor
Query Results
Debug info (Query
Proﬁle)
STATISTICS &
metadataSystem metadata
Query Explain PlanN1QL Statement
P
atent
P
ending

N1QL: Cost Based Optimizer
● N1QL Optimizer Decisions
○ Query Rewrites
○ Index selection
■ Predicate pushdown
■ Pagination pushdowns (LIMIT, OFFSET)
○ Join type and sides selection
■ Nested loop (NL) or hash join (HJ)
■ NL: decide outer and inner sides
■ HJ: decide build and probe sides
● Create the plan and query execution tree.
● CBO: Decisions will be statistics
○ BETTER DECISIONS
● Collect statistics using UPDATE STATISTICS
● Best practices for collecting statistics are documented
● ADVISE recommends UPDATE STATISTICS as well
● When statistics is absent, rule based is used.
UPDATE STATISTICS FOR `travel-sample`
(type, destinationairport, city, faa, city, country)

N1QL: Cost Based Optimizer
● Use EXPLAIN as usual to see the CBO decisions
● EXPLAIN will give you the estimated cardinality & cost
● Profile output gives you actual cardinalities as well
● USE INDEX, USE HASH will continued to be
supported to override the default decisions
"#stats": {
"#itemsOut": 168168,
"#phaseSwitches": 672677,
"execTime": "526.643124ms",
"kernTime": "227.247789ms",
"servTime": "70.269443ms"
},
"as": "route",
"cardinality": 128068.90036900375,
"cost": 70181.75740221406,
"covers": [
"cover ((`route`.`sourceairport`))",
"cover
((`route`.`destinationairport`))",
"cover ((distinct (array (`v`.`day`)
for `v` in (`route`.`schedule`) end)))",
"cover ((meta(`route`).`id`))"
],

N1QL: Index Advisor
● ADVISE on a single statement
○ Preview in 6.5, GA in 6.6.
○ Enhance for CBO in the upcoming release
● ADVISOR() on a workload
○ Preview in 6.5, GA in the upcoming release
● ADVISE uses the rules for creating indexes and gives
non-covering and covering indexes.
○ For CBO, it also advises on statistics
● ADVISOR() can be invoked manually with set of
queries or as a result of queries from completed
requests
● ADVISOR Input
○ Workload
○ Result of a query
○ Static set of queries
○ It merges the index recommendations.
○ It is 1.0. Give us feedback.
ADVISE SELECT DISTINCT route.destinationairport
FROM `travel-sample` airport JOIN `travel-sample`
route
ON airport.faa = route.sourceairport
AND route.type = "route"
WHERE airport.type = "airport"
SELECT ADVISOR({"action": "start", "response": "0s",
"duration": "1h"}) AS Collect;
SELECT ADVISOR({"action": "get", "session":
"8c41a3c6-2252-437e-ab47-0b28f29f47fb"}) AS Get;
SELECT ADVISOR((
SELECT RAW statement
FROM system:completed_requests
WHERE LOWER(statement) LIKE "select%"
AND LOWER(statement) NOT LIKE "%advis%"))
INDEX-ADVISO
R.CO
UCHB
ASE.COM

Indexing Features
5

N1QL: Data Logical Containment
Couchbase Cluster
Bucket (cxprof)
Scope (usa)
Collection
(profille)
● Indexes per collection
○ Improved performance & resource usage
● Index names have to be unique PER collection
● All the index types are still supported
○ Partial, partitioned, expression, etc
○ Partial indexes on collections with multi doc types
● Supports large number of indexes
● Tools, query, explain are all enhanced to support it
SELECT profname AS name,
uspszip AS postalcode
FROM cxprof.usa.loginfo l where l.type = "login";
CREATE INDEX i1 ON
cxprof.usa.loginfo(profname, uspszip)
PARTITION BY HASH(META().id)
GSI Indexes
on profile
GSI Indexes
on loginfo
Collection
(loginfo)

Search Features
6

FTS: Searching on Collections
Couchbase Cluster
Bucket (cxprof)
Scope (usa)
Collection
(profille)
● FTS indexes on a single collection
○ Familiar indexing and queries
○ Support for SEARCH() from N1QL & Flex indexing
● Single FTS index on data residing in multiple collections
○ Direct query from FTS
○ Simpler search
FTS Index
on profile
Collection
(loginfo)
FTS index on
documents in
gameinfo & loginfo
Collection
(gameinfo)

● Improve query performance for ad hoc queries, and provide a predictable query performance in
applications where search patterns are not predetermined. This is available in Couchbase 6.6.
● Allow applications to leverage the unique capability of Keyword analyzer in search index to support
complex queries involving combinations of AND/OR and multiple array predicates.
● Can greatly reduce the number of indexes and resource needed to support complex queries.
SELECT * FROM crm a USE INDEX (USING FTS)
WHERE a.type='activity'
AND ( a.dept = 'iA88'
OR a.region > '59416' )
AND a.priority = 'High'
AND ( a.act_date BETWEEN
'2018-01-01' AND '2018-08-31'
OR a.event.location = 'Moscone Center' )
AND ( a.account.id = 'acc100'
OR a.owner.name = 'Amanda Morrison')
Flex Index allows query service to leverage search capabilities, using standard
N1QL
N1QL: FLEX INDEX
P
atent
P
ending

DISTRIBUTED N1QL TRANSACTIONS
7

Six thousand years ago,
the Sumerians invented writing for
TRANSACTION PROCESSING

● Multi-Everything: documents, statements,
collections, scope, buckets, APIs, tools
● Ensure data integrity for application
transaction that consists of multiple
operations where multiple concurrent
activities can occur
● Support all or nothing for multi documents
N1QL DML operations
● No central coordinator for scale out
● Optimistic Concurrency Control with RYOW
● Available in via SDKs, CBQ shell, REST API
● ACID details in the next slide.
N1QL Transactions ensures database consistency when data modifications
are performed with multiple documents in a single or multiple N1QL statements
BEGIN TRANSACTION;
UPDATE customer SET balance = balance + 100 WHERE cid =
4872;
UPDATE customer SET balance = balance - 100 WHERE cid =
1924;
SELECT cid, name, balance FROM customer
WHERE cid IN [4872, 1924];
COMMIT ;
N1QL: Transactions
Patent
Pending

ACID
● ATOMIC
○ ALL or Nothing for each statement and the whole
transaction
● CONSISTENT
○ Modifications are done on the qualified set
○ Only constraint is document key unique constraint
○ Handles halloween problem via index snapshots
● ISOLATED
○ READ COMMITTED isolation
○ Monotonic atomic reads
○ Scan consistency: Request_plus by default
■ Tunable with Unbounded
○ No dirty reads
● DURABLE
○ Majority - default
○ majorityAndPersistActive
○ persistToMajority
Couchbase Cluster
Query
Query
Query
Apps
GSI
GSI
FTS
Key
Valu
e
Apps
Apps
sdk
sdk
sdk
Query
CBQ Shell
Query
Workbench

ACIDic Statements
● BEGIN TRANSACTION;
○ request to any of the query nodes. It returns txid
○ Send subsequent tx statements to same node
○ Send each statements with txid
● All DMLs are allowed. non-DMLs raise error
○ Supported: SELECT, INSERT, UPDATE,
DELETE, UPSERT, MERGE
○ Failure within a statement simply fails and rolls
back the statement. Transaction continues
● SAVEPOINT savepointname
○ Sets and names a save point
● ROLLBACK TO SAVEPOINT savepointname
○ Rolls back the changes to a named savepoint.
● ROLLBACK TRANSACTION;
○ Rollback the entire transaction
● COMMIT
○ Commits all the modifications into Couchbase
○ The whole transaction is rolled back in case of
Couchbase Cluster
Query
Query
Query
Apps
GSI
GSI
FTS
Key
Valu
e
Apps
Apps
sdk
sdk
sdk
Query
CBQ Shell
Query
Workbench

ACID for single statement
/* Simply set tximplicit to True */
UPDATE customer
SET balance = balance + 100
WHERE dept = 100;
INSERT INTO customer(KEY cid, VALUE v)
SELECT meta().id AS cid, newcust AS v
FROM newcust;
● N1QL uses optimistic concurrency control
● Caches the version of the modific documents
○ Caching the changes until commit takes memory!
○ Suitable for large number of small transactions
● When you’re trying to modify large number of docs
○ Issue the statement like before
○ Set the tximplicit parameter to True.
● When tximplicit is set to True
○ N1QL starts a new transaction
○ Each write is staged in this transaction
○ If there’s any failure, the whole operation is rolled
back

ACID Support in SDK
try {
transactions.run((ctx) -> {
// starts the transaction
TransactionGetResult cxobj = ctx.get(customer, cxid);
ctx.replace(cxobj, newaddress);
ctx.query("UPDATE customer SET balance = balance + 100 WHERE cid = 4872");
ctx.query("UPDATE customer SET balance = balance - 100 WHERE cid = 1924;");
ctx.query("SELECT cid, name, balance FROM customer WHERE cid IN [4872, 1924]");
// This call is optional - if you leave it off, the transaction
// will be committed anyway.
ctx.commit();
});
} catch (TransactionFailed e) {
// Just log the error
logger.warn("Transaction did not reach commit:");
for (LogDefer log: e.result().log().logs()) {
logger.warn(log.toString());
}
}
● The Java SDK Lambda Example
● Supports all the transactional config
● Mix and match KV and query
● More SDKs on the roadmap

● Documents qualified are done by scanning the index using the scan consistency
● RYOW: Transaction sees it’s own changes in all subsequent queries by caching the changes
○ Limit the number of updates in a multi-statement transaction
○ Use singleton transaction when necessary
● Updates are all optimistic. Avoid the hot row problem with best practices in application design
● Don’t do non-transactional writes on the documents participating in transactions
● Multi-statement transactions are lock free and coordination free. Designed for scale out
N1QL: Best Practice

WRAP
8

38
SQL Features
Triggers
T-SQL, languages
JSON, XML
SQL is English
for Relational
Database
SQL Invented by Don
Chamberlin &
Raymond Boyce at
IBM
N1QL, based on
SQL, is English for
JSON
Gerald Sangudi at
Couchbase
SQL
Instance
Database
Schema
Table
Row
Column
N1QL
Cluster
Bucket
Scope
Collection
Document
Attribute
SQL
Tuples
N1QL STMT
CREATE BUCKET
CREATE INDEX
None
SELECT
INSERT
UPDATE
DELETE
MERGE
Subqueries
JOIN
GROUP BY
ORDER BY
OFFSET, LIMIT
EXPLAIN
PREPARE
EXECUTE
GRANT ROLE
REVOKE ROLE
INFER
PREPARE
EXECUTE
ADVISE (6.6)
ADVISOR
UPDATE-STATISTICS
CREATE FUNCTION
BEGIN WORK
COMMIT WORK
SAVEPOINT
ROLLBACK
EXECUTE
FUNCTION
Set of
JSON
N1QL Model
Set of
JSON
SQL
Tooling
ODBC, JDBC, .NET
Hibernate
BI Tools
erwin
TOAD
SQL Indexes
Primary Key
Secondary Key
Composite
Range Partitioned
Expression
Functional
Spatial
Search
SQL Logic
3 valued logic
N1QL
Tooling
ODBC,JDBC, .NET
Couchbase SDKs
Spring
BI Tools
erwin
SQL
TRANSACTIONS
ACID
Multi-Statement
Multi-table
Mullti-Database
Savepoints
Commit
2 phase locking
Single node
N1QL
TRANSACTIONS
ACID
Multi-Statement
Mullti-collection
Multi-Bucket
Savepoints
Commit
OCC (Optimistic)
READ COMMITTED
Distributed tx
SQL Datatypes
Numeric
Decimal
Character
Date Time
Timezone
BLOB
Spatial
JSON
N1QL
Datatype
Numeric
Boolean
Character
Array
Object
Null
JSON
SQL
Optimizer
Rule Based
Cost Based
Index Selection
Join reordering
Query Rewrites
N1QL
Optimizer
Rule Based
Cost Based
Index Selection
Join as specified
Query Rewrites(Ltd)
NL, Hash join
N1QL
Engine
(SCALE UP &
SCALE OUT)
SQL STMT
CREATE TABLE
CREATE INDEX
ALTER TABLE
SELECT
INSERT
UPDATE
DELETE
MERGE
Subqueries
JOIN
GROUP BY
ORDER BY
OFFSET, LIMIT
EXPLAIN PLAN
PREPARE
EXECUTE
GRANT
REVOKE
DESCRIBE
PREPARE
EXECUTE
TRUNCATE
N1QL
Input and Output:
Set(s) of JSON
Tuples
SQL Model
Set of
Tuples
SQL Engine
(SCALE UP)
N1QL Features
Eventing Service
Defacto JSON
N1QL Logic
4 valued logic
TRUE, FALSE,
N1QL
Indexes
Primary
Secondary
Composite
Partial
Expression
Functional
Array Index
Search
Flex Index(6.6)

THANK YOU

N1QL New Features in couchbase 7.0

More Related Content

What's hot (20)

Similar to N1QL New Features in couchbase 7.0 (20)

More from Keshav Murthy (20)

Recently uploaded (20)

N1QL New Features in couchbase 7.0