Understanding Datasets#

DEPRECATED

Apperate and its features (including this feature) are deprecated.

Datasets are schema-based, time series-optimized database tables that have auto-generated, permissioned APIs and API docs. You can query and operate on datasets in a codeless fasion in the console or programatically using Apperate’s GET /data method or our open-source client libraries.

Apperate validates incoming data against the dataset’s schema, giving you further confidence in your data. Each property (think database column) has a type, is required (or not), and allows null (or not).

Indexing with Unique Index#

Each dataset has a Unique Index composed of a primary index, secondary index (optional), and date index. This index determines how the dataset records are identified, deduplicated, partitioned, and stored.

Unique Index = primary + secondary + date

primary and secondary (optional) indexes: Set these to any property. Indexes should ideally be assigned to properties that signify an important identifier and/or category for the dataset.

date index: Set this to any date property or leave it unset to use record creation timestamps. The property mapped to the date index is the primary mechanism by which data is time-sorted.

Querying on Key, Subkey, and Date#

The Data API leverages the Unique Index components. It nicknames them as follows …

Key (primary index)
Subkey (secondary index)
Date (date index)

Query Examples#

Here’s an example of mapping for a company news dataset.

Nickname	Property	Example
key	`symbol`	A stock symbol `AAPL`
subkey	`source`	Information source `CNN`
date	`timestamp`	The record’s date/time `2019-02-20`

Here’s a URL format that queries on all three components:

BASE_URL/data/WORKSPACE/DATASET/KEY/SUBKEY?token=TOKEN&on=DATE

Since your base URL, dataset, and workspace are already known, you can concentrate on plugging key (AAPL), subkey (CNN), and date (2019-02-20) values into your GET /data requests.

Here are some example queries.

Get the AAPL news record from CNN on 2019-02-20:

https://myworkspace.iex.cloud/v1/data/MYWORKSPACE/MY_DATASET/AAPL/CNN?token=TOKEN&on=2019-02-20

Get all AAPL news records from any source on 2019-02-20:

https://myworkspace.iex.cloud/v1/data/MYWORKSPACE/MY_DATASET/AAPL?token=TOKEN&on=2019-02-20

Get all AAPL news records from CNN:

https://myworkspace.iex.cloud/v1/data/MYWORKSPACE/MY_DATASET/AAPL/CNN?token=TOKEN

Get all AAPL news records from any source:

https://myworkspace.iex.cloud/v1/data/MYWORKSPACE/MY_DATASET/AAPL?token=TOKEN

Normalizing with SmartLinks#

You can opt-in (SmartLink) a primary or secondary index property with IEX Cloud’s metadata graph. A SmartLink associates the property’s values with equivalent values from any of Apperate’s supported financial identifier types.

The image below shows a dataset’s symbol property opted in to SmartLinks.

For example, if you SmartLink a property that holds financial symbols that use FIGI, you can match the symbols using equivalent symbols that use other financial identifier types, such as ISIN or CUSIP.

By SmartLinking a property you can:

Enrich your dataset by joining to any and all of IEX Cloud’s equities data
Join this dataset to any other dataset you have in IEX Cloud that is opted-in to IEX Cloud’s metadata graph
Ingest and query for data in this dataset using any of IEX Cloud’s supported financial identifiers

Permissioning with Tokens#

Your datasets are private by default. You can share a dataset’s API with other parties by giving them a public API token that has access to the dataset.

What’s Next#

Data Model Examples demonstrates using dataset schemas and the Unique Index to represent various data models.

Updating Dataset Schemas shows you how to update dataset schemas and explains how updates work.

Search Data: These articles show how to query data.