dedupe.md

dedupe

The dedupe processor is used to dedupe an array of DataEntities or an array of DataWindows by a given field. If no field is configured then it will attempt to dedupe based off the _key metadata property. This processor can also track dates of duplicate records so that the resulting unique record has either the oldest or newest date for the date field based on the adjust_time parameter.

Usage

Dedupe records based on a field

Example of a job using the dedupe processor

{
    "name" : "testing",
    "workers" : 1,
    "slicers" : 1,
    "lifecycle" : "once",
    "assets" : [
        "standard"
    ],
    "operations" : [
        {
            "_op": "test-reader"
        },
        {
             "_op": "dedupe",
             "field": "name"
        }
    ]
}

Output from example job

const data = [
    { id: 1, name: 'roy' },
    { id: 2, name: 'roy' },
    { id: 2, name: 'bob' },
    { id: 2, name: 'roy' },
    { id: 3, name: 'bob' },
    { id: 3, name: 'mel' }
]

const results = await processor.run(data);

results === [
    { id: 1, name: 'roy' },
    { id: 2, name: 'bob' },
    { id: 3, name: 'mel' }
];

Dedupe records based on the _key metadata

Example of a job using the _key in the metadata

{
    "name" : "testing",
    "workers" : 1,
    "slicers" : 1,
    "lifecycle" : "once",
    "assets" : [
        "standard"
    ],
    "operations" : [
        {
            "_op": "test-reader"
        },
        {
             "_op": "dedupe"
        }
    ]
}

Output from example job

const data = [
    DataEntity.make({ id: 1, name: 'roy' }, { _key: 1 }),
    DataEntity.make({ id: 2, name: 'roy' }, { _key: 2 }),
    DataEntity.make({ id: 2, name: 'bob' }, { _key: 2 }),
    DataEntity.make({ id: 2, name: 'roy' }, { _key: 2 }),
    DataEntity.make({ id: 3, name: 'bob' }, { _key: 3 }),
    DataEntity.make({ id: 3, name: 'mel' }, { _key: 3 }),
];

const results = await processor.run(data);

results === [
    { id: 1, name: 'roy' },
    { id: 2, name: 'roy' },
    { id: 3, name: 'bob' }
];

Dedupe records and track time

Example of a job using the dedupe processor and tracking the oldest date of the first_seen field as well as the newest date of the last_seen field.

{
    "name" : "testing",
    "workers" : 1,
    "slicers" : 1,
    "lifecycle" : "once",
    "assets" : [
        "standard"
    ],
    "operations" : [
        {
            "_op": "test-reader"
        },
        {
            "_op": "dedupe",
            "field": "name",
            "adjust_time": [
                { "field": "first_seen", "preference": "oldest" },
                { "field": "last_seen", "preference": "newest" }
            ]
        }
    ]
}

Output of example job

const data = [
    {
        id: 1,
        name: 'roy',
        first_seen: '2019-05-07T20:01:00.000Z',
        last_seen: '2019-05-07T20:01:00.000Z'
    },
    {
        id: 1,
        name: 'roy',
        first_seen: '2019-05-07T20:02:00.000Z',
        last_seen: '2019-05-07T20:02:00.000Z'
    },
    {
        id: 1,
        name: 'roy',
        first_seen: '2019-05-07T20:04:00.000Z',
        last_seen: '2019-05-07T20:04:00.000Z'
    },
    {
        id: 2,
        name: 'bob',
        first_seen: '2019-05-07T20:02:00.000Z',
        last_seen: '2019-05-07T20:02:00.000Z'
    },
    {
        id: 1,
        name: 'roy',
        first_seen: '2019-05-07T20:10:00.000Z',
        last_seen: '2019-05-07T20:10:00.000Z'
    },
    {
        id: 2,
        name: 'bob',
        first_seen: '2019-05-07T20:04:00.000Z',
        last_seen: '2019-05-07T20:04:00.000Z'
    },
    {
        id: 3,
        name: 'mel',
        first_seen: '2019-05-07T20:04:00.000Z',
        last_seen: '2019-05-07T20:04:00.000Z'
        },
    {
        id: 1,
        name: 'roy',
        first_seen: '2019-05-07T19:02:00.000Z',
        last_seen: '2019-05-07T19:02:00.000Z'
    },
    {
        id: 1,
        name: 'roy',
        first_seen: '2019-05-07T20:08:00.000Z',
        last_seen: '2019-05-07T20:08:00.000Z'
    },
    {
        id: 2,
        name: 'bob',
        first_seen: '2019-05-07T20:08:00.000Z',
        last_seen: '2019-05-07T20:08:00.000Z'
    },
    {
        id: 3,
        name: 'mel',
        first_seen: '2019-05-07T20:01:00.000Z',
        last_seen: '2019-05-07T20:01:00.000Z'
    }
];

const results = await processor.run(data);

results === [
    {
        id: 1,
        name: 'roy',
        first_seen: '2019-05-07T19:02:00.000Z',
        last_seen: '2019-05-07T20:10:00.000Z'
    },
    {
        id: 2,
        name: 'bob',
        first_seen: '2019-05-07T20:02:00.000Z',
        last_seen: '2019-05-07T20:08:00.000Z'
    },
    {
        id: 3,
        name: 'mel',
        first_seen: '2019-05-07T20:01:00.000Z',
        last_seen: '2019-05-07T20:04:00.000Z'
    }
]

Parameters

Configuration	Description	Type	Notes
_op	Name of operation, it must reflect the exact name of the file	String	required
field	field to dedupe records on	String	optional, defaults to _key metadata value
adjust_time	Requires an array of objects with field and preference properties. Preference should be set to oldest or newest.	Array of Objects	optional, defaults to []