Skip to content 🧪

Pages table

httparchive.all.pages is a partitioned and clustered table containing one row per page tested in the HTTP Archive. Pages are tested on a monthly basis and as of April 2022, both the root page and one secondary page are tested.

Example queries

Here are some common operations you can perform with the pages table.

Get the median page weight

/* This query will process 1.12 GB when run. */
WITH pages AS (
    CAST(JSON_VALUE(summary, '$.bytesTotal') AS INT64) AS page_weight
    `httparchive.all.pages` TABLESAMPLE SYSTEM (1 PERCENT)
    date = '2023-05-01'

  APPROX_QUANTILES(page_weight, 1000)[OFFSET(500)] AS median_page_weight

This query uses the APPROX_QUANTILES function to calculate the median page weight for each client type as of May 2023.

The bytesTotal property of the summary object represents the total number of bytes loaded on the page. This value is stored as a JSON-encoded string, so we use JSON_VALUE to extract it and CAST to convert it to an integer.

We’re also using the WITH clause here to create a temporary table called pages, which is then fed into the main query below. This makes the query a bit easier to read.

Also note that for demonstration purposes, this query processes a 1% sample of the httparchive.all.pages table. This reduces the amount of data processed by the query, which can help reduce costs. But note that the results will be less accurate than if you ran the query on the full table.


Field nameTypeDescription
dateDATEYYYY-MM-DD format of the HTTP Archive monthly crawl
clientSTRINGTest environment: 'desktop' or 'mobile'
pageSTRINGThe URL of the page being tested
is_root_pageBOOLEANWhether the page is the root of the origin
root_pageSTRINGThe URL of the root page being tested, the origin followed by /
rankINTEGERSite popularity rank, from CrUX
wptidSTRINGID of the WebPageTest results
payloadHARJSON-encoded WebPageTest results for the page
summarySTRINGJSON-encoded summarization of the page-level data
custom_metricsSTRINGJSON-encoded test results of the custom metrics
lighthouseLighthouseJSON-encoded Lighthouse report
featuresARRAY<Feature>Blink features detected at runtime
technologiesARRAY<Technology>Technologies detected at runtime
metadataSTRINGAdditional metadata about the test


This field is required for all queries over the pages table.

YYYY-MM-DD format of the HTTP Archive monthly crawl.

Example: date = '2023-06-01'


Test environment: 'desktop' or 'mobile'.


The URL of the page being tested.

Example: page = ''


Whether the page is the root of the origin.


The URL of the root page being tested, the origin followed by /.

Example: root_page = ''


Site popularity rank, from CrUX


ID of the WebPageTest results, for example wptid = '230509_Dx20W_FMHK5'.

The ID encodes the date of the test at the start in YYMMDD format. The date is followed by an underscore and a D or M character indicating whether it was a desktop or mobile test. The rest of the ID is randomly generated. In the example above we can tell that the page was tested on May 9, 2023, and that it was a desktop test.

You can view the WebPageTest results in the browser by visiting<wptid>/, eg This is HTTP Archive’s own private instance of WebPageTest, which is required to view any of the results.


JSON-encoded WebPageTest results for the page.

For a full example value, see payload.json.

See the har reference for more details.


JSON-encoded summarization of the page-level data


JSON-encoded test results of the custom metrics.

See the custom metrics reference for more details.


JSON-encoded Lighthouse report.

See the lighthouse reference for more details.


Blink features detected at runtime (see


Technologies detected at runtime (see

See the technology reference for more details.


Additional metadata about the test

See the metadata reference for more details.