Importing items incrementally
After you create an Items dataset, you can incrementally import one or more new items into the dataset. Incrementally importing items allows you to keep your Items dataset current as your catalog grows. If you have a large amount of new items, we recommend that you first import data in bulk and then import item data incrementally as necessary. See Importing bulk records.
You can use the Amazon Personalize console, the Amazon Command Line Interface (Amazon CLI), or Amazon SDKs to import
items. If you import an item with the same itemId
as an item that's already in your
Items dataset, Amazon Personalize replaces it with the new item. You can import up to 10 items at a
time.
For information about how Amazon Personalize updates filters for new records and how new records influence recommendations, see Importing records incrementally.
Topics
Importing items incrementally (console)
You can import up to 10 items to an Items dataset at a time. This procedure assumes that you have already created an Items dataset. For information about creating datasets, see Step 2: Creating a dataset and a schema.
To import items incrementally (console)
-
Open the Amazon Personalize console at https://console.amazonaws.cn/personalize/home
and sign in to your account. -
On the Dataset groups page, choose the dataset group with the Items dataset that you want to import the items to.
-
In the navigation pane, choose Datasets.
-
On the Datasets page, choose the Items dataset.
-
At the top right of the dataset details page, choose Modify dataset, and then choose Create record.
-
In Create item record(s) page, for Record input, enter the item details in JSON format. The item's field names and values must match the schema you used when you created the Items dataset. Amazon Personalize provides a JSON template with field names and data types from this schema.
-
Choose Create record(s). In Response, the result of the import is listed and a success or failure message is displayed.
Importing items incrementally (Amazon CLI)
Add one or more items to your Items dataset using the PutItems operation. You can import up to 10 items with a single
PutItems
call. This section assumes that you have already created an Items dataset. For information
about creating datasets, see Step 2: Creating a dataset and a schema.
Use the following put-items
command to add one or more items with the Amazon CLI.
Replace dataset arn
with the Amazon Resource Name (ARN) of your dataset and
item Id
with the ID of the item. If an item with the same itemId
is already in your Items dataset, Amazon Personalize replaces it with the new one.
For properties
, for each field in your Items dataset, replace the
propertyName
with the field name from your schema in camel case. For example,
GENRES would be genres
and CREATION_TIMESTAMP would be creationTimestamp. Replace item data
with the data
for the item. CREATION_TIMESTAMP
data must be in Unix epoch time format and in seconds. For categorical string data, to include
multiple categories for a single property, separate each category with a pipe
(|
). For example \"Horror|Action\"
.
aws personalize-events put-items \ --dataset-arn
dataset arn
\ --items '[{ "itemId": "item Id
", "properties": "{\"propertyName
\": "\item data
\"}" }, { "itemId": "item Id
", "properties": "{\"propertyName
\": "\item data
\"}" }]'
Importing items incrementally (Amazon SDKs)
Add one or more items to your Items dataset using the PutItems operation. You can import up to 10 items with a single
PutItems
call. This section assumes that you have already created an Items dataset. For information
about creating datasets, see Step 2: Creating a dataset and a schema.
The following code shows how to add one or more items to your Items dataset with the Amazon SDK for Python (Boto3) or the Amazon SDK for Java 2.x.