All Products
Search

This Product

    OpenSearch:Import data

    Document Center

    OpenSearch:Import data

    Last Updated:Apr 01, 2026

    OpenSearch LLM-based Conversational Search Edition supports four data import methods. Choose based on your data source:

    MethodBest for
    Upload filesLocal files — documents, spreadsheets, PDFs
    Import from web pagesA known set of specific URLs
    Import from a websiteAn entire site or section, crawled automatically
    Import tablesStructured tabular data for table-based Q&A

    Prerequisites

    Before you begin, ensure that you have:

    • An active OpenSearch LLM-based Conversational Search Edition instance

    • Access to the OpenSearch console

    Get to the data configuration page

    1. Log on to the OpenSearch console.

    2. In the top navigation bar, select the region where your instance resides. In the upper-left corner, select OpenSearch LLM-Based Conversational Search Edition.

    3. On the Instance Management page, find your instance and click Manage in the Actions column.

    4. In the left pane, choose Configuration Center > Data Configuration, then select a data import method.

    Data structure

    The primary table uses a fixed schema with the following fields:

    FieldTypeRequiredDescription
    idLITERALYesUnique document ID
    contentTEXTYesDocument content
    titleTEXTNoDocument title
    categoryLITERAL_ARRAYNoDocument category. Separate multiple values with commas (,). See Custom parameters.
    urlLITERALNoDocument URL
    scoreFLOATNoDocument score. Affects retrieval ranking. See Document retrieval parameters.
    timestampINTNoDocument freshness indicator. Affects retrieval ranking. See Document retrieval parameters.

    Upload files

    Click Import File. The Import File panel opens.

    image

    Supported formats and limits

    Data typeSupported formatsMax file sizeNotes
    Unstructured dataDOC, DOCX, PDF, HTML, TXT, PPT, PPTX128 MB per fileMultiple files allowed. If a Word file contains too many images, we recommend that you convert the file format to PDF to speed up the upload.
    Structured dataJSON, Excel (UTF-8 encoded)128 MB per fileMultiple files allowed. Click Data Sample to view the expected format.

    Structured data naming constraints:

    • Table name: letters, digits, and underscores (_); max 20 characters

    • Field name: letters and underscores (_); cannot start with _; max 30 characters

    • Excel fields: max 30 fields per file are imported and queried; excess fields are ignored

    The total size of structured data uploaded in a single operation cannot exceed 2 MB.

    To push multiple documents at once using the API, call PushKnowledgeDocuments.

    Import data from web pages

    Use this method when you have a specific list of URLs to import.

    Click Web Page URL Import. On the Web Page Import tab, enter the URLs one per line, then click Import.

    网页链接.png

    Import data from a website

    Use this method to crawl and import an entire website or a section of one.

    1. Click Web Page URL Import, then click the Website Import tab.

    2. Click Create Task.

    3. In the Create Task dialog box, configure the required parameters, then click OK.

    新建任务.png

    Required parameters:

    ParameterDescription
    Website URLRoot URL of the site to crawl
    CategoryCategory assigned to all imported content

    Optional parameters:

    ParameterDescriptionExample
    URL FilteringRegular expression controlling which URLs are crawled. The default rule is a regular expression that starts with the website URL.If the website URL is http://www.abc.com/, the default filter is http://www\.abc\.com/.*
    XPath SelectorLimits imported content to specific HTML elements//div imports content inside <div> elements
    CSS SelectorLimits imported content to elements matching a CSS selectordiv.content imports <div class="content"> elements
    URLs ending with .png, .jpg, or .jpeg are not supported.

    1. After the task is configured, click OK. The system shows the number of crawled pages. Click OK again to start the import.

    导入完成.png

    Import tables

    For table-based conversational search, see Implement table-based conversational search.

    Query and manage documents

    After upload, the Data Query section shows the total document count. Depending on the volume of data, there may be a wait before documents are queryable. To test Q&A against your data, go to the Q&A test page.

    To look up a specific document, select id from the drop-down list, enter the document ID, and click the Search icon.

    View a document

    Select id, enter the document ID, click the Search icon, then view the document details.

    image.png

    Edit a document

    Select id, enter the document ID, click the Search icon, then click Edit in the Actions column to modify editable fields.

    image.png

    Delete a document

    Select id, enter the document ID, click the Search icon, then choose More > Delete in the Actions column.

    If storage capacity is insufficient, deletion fails. Expand your storage capacity first.
    image.png

    Usage notes

    • Each document's primary key must be unique. If two documents share the same primary key, the newer document overwrites the earlier one.

    • The size of the structured data that you upload at a time cannot exceed 2 MB.

    • The size of an unstructured data file that you upload cannot exceed 128 MB.

    • After upload, the time before documents become queryable depends on the volume of data being indexed.