Warning: This page is about Google's older APIs, the Google Data APIs; it's relevant only to the APIs that are listed in the Google Data APIs directory, many of which have been replaced with newer APIs. For information about a specific new API, see the new API's documentation. For information about authorizing requests with a newer API, see Google Accounts Authentication and Authorization.
This document describes the Google Data Protocol used by many Google APIs, including information about what a query looks like, what results look like, and so on.
- Protocol details
- Additional resources
This document is intended for anyone wanting to understand the details of the XML format and protocol used by the APIs that implement the Google Data Protocol.
If you just want to write code that uses one of these APIs, then you don't need to know these details; instead, you can use the language-specific client libraries.
But if you want to understand the protocol, read this document. For example, you may want to read this document to help you with any of the following tasks:
- evaluating the Google Data Protocol architecture
- coding using the protocol without using the provided client libraries
- writing a client library in a new language
This document assumes that you understand the basics of XML, namespaces, syndicated feeds, and the
DELETE requests in HTTP, as well as HTTP's concept of a "resource." For more information about those things, see the Additional resources section of this document.
This document doesn't rely on any particular programming language; you can send and receive Google Data Protocol messages using any programming language that lets you issue HTTP requests and parse XML-based responses.
This section describes the Google Data Protocol document format and query syntax.
The Google Data Protocol and Atom share the same basic data model: a container that holds both some global data and any number of entries. For each protocol, the format is defined by a base schema, but it can be extended using foreign namespaces.
Atom is the default format for the Google Data Protocol. To request a response in another format, use the
alt query parameter; for more information, see Query requests.
Note: Most Google Data Protocol feeds in Atom format use the Atom namespace as the default namespace by specifying an
xmlns attribute on the feed element, as seen in the examples given in Protocol Basics. Thus, the examples in this document don't explicitly specify
atom: for elements in an Atom-format feed.
The following tables show the Atom representation of the elements of the schema. All data not mentioned in these tables is treated as plain XML. Unless indicated otherwise, the XML elements in a given column are in the Atom namespace.
Note: This summary uses standard XPath notation: in particular, slashes show the element hierarchy, and an @ sign indicates an attribute of an element.
In each of the following tables, the highlighted items are required.
The following table shows the elements of a Google Data Protocol feed:
|Feed Schema Item||Atom Representation|
|Feed HTML Link||
(Required in certain cases; see Atom specification.)
|Feed Last Update Date||
(RFC 3339 format)
|Feed Category Scheme||
The following table shows the elements of a Google Data Protocol search-results feed. Note that the protocol exposes some of the OpenSearch 1.1 Response elements in search-results feeds.
|Search Result Feed Schema Item||Atom Representation|
|Number of Search Results||
|Search Result Start Index||
|Number of Search Results Per Page||
The following table shows the elements of a Google Data Protocol entry:
|Entry Schema Item||Atom Representation|
(Required in certain cases; see Atom specification.)
(If no content element, then entry must contain at least one
(Required in certain cases; see Atom specification.)
|Entry Category Scheme||
|Entry Publication Date||
|Entry Update Date||
This section describes how to use the query system.
Query model design tenets
The query model is intentionally very simple. The basic tenets are:
- Queries are expressed as HTTP URIs, rather than as HTTP headers or as part of the payload. One benefit of this approach is that you can link to a query.
- Predicates are scoped to a single item. Thus, there's no way to send a correlation query such as "find all emails from people who sent me at least 10 emails today."
- The set of properties that queries can predicate on is very limited; most queries are simply full text search queries.
- Result ordering is up to the implementation.
- The protocol is naturally extensible. If you want to expose additional predicates or sorting in your service, you can do so easily through the introduction of new parameters.
A client queries a Google service by issuing an HTTP
GET request. The query URI consists of the resource's URI (called FeedURI in Atom) followed by query parameters. Most query parameters are represented as traditional
?name=value[&...] URL parameters. Category parameters are handled differently; see below.
For example, if the FeedURI is
http://www.example.com/feeds/jo, then you might send a query with the following URI:
The Google Data Protocol supports HTTP Conditional
GET. APIs that implement the protocol set the Last-Modified response header based upon the value of the
<atom:updated> element in the returned feed or entry. A client can send this value back as the value of the If-Modified-Since request header to avoid retrieving the content again if it hasn't changed. If the content hasn't changed since the If-Modified-Since time, then the service returns a 304 (Not Modified) HTTP response.
APIs that implement the Google Data Protocol must support
alt queries; support for other parameters is optional. Passing a standard parameter not understood by a given service results in a
403 Forbidden response. Passing an unsupported nonstandard parameter results in a
400 Bad Request response. For information on other status codes, see the HTTP status codes section of this document.
The standard query parameters are summarized in the following table. All parameter values need to be URL encoded.
||Alternative representation type||
||Category query filter||
||Category query filter||
|entryID||ID of a specific entry to be retrieved||
||Maximum number of results to be retrieved||For any service that has a default
||Returns an XML response with identations and line breaks||
||Bounds on the entry publication date||
||Full-text query string||
||1-based index of the first result to be retrieved||
||Strict query parameter checking||
||Bounds on the entry update date||
About category queries
We decided to provide a slightly unusual format for category queries. Instead of requiring a query like the following:
we made it possible to use:
This approach identifies a resource without using query parameters, and it produces cleaner URIs. We chose this approach for categories because we think that category queries will be among the most common queries.
The drawback to this approach is that we require you to use
/-/ in this type of category queries, so that services can distinguish category queries from other resource URIs, such as
Queries return an Atom feed, an Atom entry, or an RSS feed, depending on the request parameters.
Query results contain the following OpenSearch elements directly under the
<feed> element or the
<channel> element (depending on whether results are Atom or RSS):
- The total number of search results for the query (not necessarily all present in the results feed).
- The 1-based index of the first result.
- The maximum number of items that appear on one page. This allows clients to generate direct links to any set of subsequent pages. However, for a possible pitfall in using this number, see the note regarding
start-indexin the table in the Query requests section.
The Atom response feed and entries may also include any of the following Atom and Data API elements (as well as others listed in the Atom specification):
<link rel="http://schemas.google.com/g/2005#feed" type="application/atom+xml" href="..."/>
- Specifies the URI where the complete Atom feed can be retrieved.
<link rel="http://schemas.google.com/g/2005#post" type="application/atom+xml" href="..."/>
- Specifies the Atom feed's PostURI (where new entries can be posted).
<link rel="self" type="..." href="..."/>
- Contains the URI of this resource. The value of the
typeattribute depends on the requested format. If no data changes in the interim, sending another GET to this URI returns the same response.
<link rel="previous" type="application/atom+xml" href="..."/>
- Specifies the URI of the previous chunk of this query result set, if it is chunked.
<link rel="next" type="application/atom+xml" href="..."/>
- Specifies the URI of the next chunk of this query result set, if it is chunked.
<link rel="edit" type="application/atom+xml" href="..."/>
- Specifies the Atom entry's EditURI (where you send an updated entry).
Here's a sample response body, in response to a search query:
<?xml version="1.0" encoding="UTF-8"?> <feed xmlns="http://www.w3.org/2005/Atom" xmlns:openSearch="http://a9.com/-/spec/opensearch/1.1/" xmlns:gd='http://schemas.google.com/g/2005' gd:etag='W/"C0QBRXcycSp7ImA9WxRVFUk."'> <id>http://www.example.com/feed/1234.1/posts/full</id> <updated>2005-09-16T00:42:06Z</updated> <title type="text">Books and Romance with Jo and Liz</title> <link rel="alternate" type="text/html" href="http://www.example.net/"/> <link rel="http://schemas.google.com/g/2005#feed" type="application/atom+xml" href="http://www.example.com/feed/1234.1/posts/full"/> <link rel="http://schemas.google.com/g/2005#post" type="application/atom+xml" href="http://www.example.com/feed/1234.1/posts/full"/> <link rel="self" type="application/atom+xml" href="http://www.example.com/feed/1234.1/posts/full"/> <author> <name>Elizabeth Bennet</name> <email>email@example.com</email> </author> <generator version="1.0" uri="http://www.example.com">Example Generator Engine</generator> <openSearch:totalResults>2</openSearch:totalResults> <openSearch:startIndex>0</openSearch:startIndex> <entry gd:etag='W/"C0QBRXcycSp7ImA9WxRVGUo."'> <id>http://www.example.com/feed/1234.1/posts/full/4521614025009481151</id> <published>2005-01-09T08:00:00Z</published> <updated>2005-01-09T08:00:00Z</updated> <category scheme="http://www.example.com/type" term="blog.post"/> <title type="text">This is the title of entry 1009</title> <content type="xhtml"> <div xmlns="http://www.w3.org/1999/xhtml">This is the entry body of entry 1009</div> </content> <link rel="alternate" type="text/html" href="http://www.example.com/posturl"/> <link rel="edit" type="application/atom+xml" href="http://www.example.com/feed/1234.1/posts/full/4521614025009481151"/> <author> <name>Elizabeth Bennet</name> <email>firstname.lastname@example.org</email> </author> </entry> <entry gd:etag='W/"C0QBRXrurSp7ImA9WxRVGUo."'> <id>http://www.example.com/feed/1234.1/posts/full/3067545004648931569</id> <published>2005-01-07T08:00:00Z</published> <updated>2005-01-07T08:02:00Z</updated> <category scheme="http://www.example.com/type" term="blog.post"/> <title type="text">This is the title of entry 1007</title> <content type="xhtml"> <div xmlns="http://www.w3.org/1999/xhtml">This is the entry body of entry 1007</div> </content> <link rel="alternate" type="text/html" href="http://www.example.com/posturl"/> <link rel="edit" type="application/atom+xml" href="http://www.example.com/feed/1234.1/posts/full/3067545004648931569"/> <author> <name>Elizabeth Bennet</name> <email>email@example.com</email> </author> </entry> </feed>
If the requested feed is in the Atom format, if no query parameters are specified, and if the result doesn't contain all the entries, the following element is inserted into the top-level feed:
<link rel="next" type="application/atom+xml" href="..."/>. It points to a feed containing the next set of entries. Subsequent sets contain a corresponding
<link rel="previous" type="application/atom+xml" href="..."/> element. By following all the next links, a client can retrieve all entries from a feed.
HTTP status codes
The following table describes what various HTTP status codes mean in the context of the Data APIs.
|200 OK||No error.|
|201 CREATED||Creation of a resource was successful.|
|304 NOT MODIFIED||The resource hasn't changed since the time specified in the request's If-Modified-Since header.|
|400 BAD REQUEST||Invalid request URI or header, or unsupported nonstandard parameter.|
|401 UNAUTHORIZED||Authorization required.|
|403 FORBIDDEN||Unsupported standard parameter, or authentication or authorization failed.|
|404 NOT FOUND||Resource (such as a feed or entry) not found.|
|409 CONFLICT||Specified version number doesn't match resource's latest version number.|
|410 GONE||Requested change history is no longer available on the server. Refer to service-specific documentation for more details.|
|500 INTERNAL SERVER ERROR||Internal error. This is the default code that is used for all unrecognized server errors.|
Resource versioning (ETags)
Sometimes you need to be able to refer to a specific version of a particular entry.
This is important in two cases in particular:
- Doing a "conditional retrieval," in which your client requests an entry, and the server sends the entry only if it has changed since the last time the client requested it.
- Ensuring that multiple clients don't inadvertently overwrite one another's changes. The Data APIs do this by making updates and deletes fail if the client specifies an old version identifier for the entry.
The Google Data APIs handle both of these cases using ETags, a standard part of HTTP.
An ETag is an identifier that specifies a particular version of a particular entry. The server attaches an ETag to entry and feed elements that it sends to clients. When an entry or feed changes, its ETag changes as well.
The Google Data APIs provide ETags in two places: in an
ETag HTTP header, and in a
gd:etag attribute of
In the Google Data APIs, an ETag is usually a string of letters and numbers, sometimes also including hyphens and periods; the string is usually enclosed in quotation marks. (The quotation marks are part of the ETag.) For example, here's an ETag from a Data API entry:
There are two kinds of ETags: strong and weak. Strong ETags identify a specific version of a specific entry, and can be used to avoid overwriting other clients' changes. Weak ETags, in the context of the Google Data APIs, are used only for conditional retrieval. A weak ETag always starts with
W/. For example:
Not all Google Data APIs support strong ETags. For those that do, the strong ETags are used only for entries; ETags on feeds are always weak.
Here's an example of a feed (including some of the HTTP headers) retrieved from a service that supports strong ETags:
GData-Version: 2.0 ETag: W/"C0QBRXcycSp7ImA9WxRVFUk." ... <?xml version='1.0' encoding='utf-8'?> <feed xmlns='http://www.w3.org/2005/Atom' xmlns:gd='http://schemas.google.com/g/2005' gd:etag='W/"C0QBRXcycSp7ImA9WxRVFUk."'> ... <entry gd:etag='"CUUEQX47eCp7ImA9WxRVEkQ."'> ... </entry> </feed>
The client libraries that support version 2 of the Data APIs handle ETags for you, transparently. The following information is for clients that don't use client libraries, and for readers interested in how versioning is handled at the protocol level.
Note: For information about the resource-versioning system used in version 1.0 of the Data APIs, see 1.0 reference guide.
If you want to retrieve an entry that you've retrieved before, you can improve efficiency by telling the server to send the entry only if it has changed since the last time you retrieved it.
To do this sort of conditional retrieval, send an HTTP
GET request that includes an HTTP
If-None-Match header. In the header, specify the entry's ETag.
Here's an example of an
When the server receives this request, it checks to see whether the entry that you requested has the same ETag as the ETag you specified. If the ETags match, then the entry hasn't changed, and the server returns an HTTP 304
Not Modified status code.
If the ETags don't match, then the entry has been modified since the last time you requested it, and the server returns the entry.
The easiest way to avoid overwriting another client's changes is for the server to make sure that when your client sends an updated entry, the version of the entry that your client started with is the same as the current version stored by the server. If a second client makes an update before your client does, then your client's update is denied, because your client is no longer basing its modifications on the latest version.
When your client retrieves data from a service that supports strong ETags, each entry has an ETag that acts as a unique version identifier for that version of that entry.
Note: Updating using ETags works only with strong ETags. For services that supply weak ETags, all updates succeed, regardless of whether someone else has updated the entry since you retrieved it; the newest update always overwrites any other prior updates. So don't send weak ETags when updating or deleting; you'll get an error message if you do.
So when your client sends an update to a strong-ETags service, it needs to specify what version of the entry it's updating. There are two ways to do that:
- Use an
- Use the
gd:etagattribute in the
We recommend the
If-Match approach where possible.
To update an entry using
If-Match, start by acquiring the entry you're updating. Make any desired changes to the entry, then create a new
PUT request containing the modified entry. (For details of URLs to use, see service-specific documentation.)
Before sending the
PUT, add an HTTP
If-Match header containing the ETag from the original entry:
Then send the
If the update succeeds, then the server returns an HTTP
200 OK status code, and a copy of the updated entry.
If the update fails because the ETag you specified doesn't match the current ETag on the entry (which implies that the entry has changed on the server since you last retrieved it), then the server returns an HTTP
412 Precondition Failed status code.
If you can't easily write HTTP headers, or have some other reason to avoid using the
If-Match header, then you can use the
gd:etag attribute instead.
If you don't send an
If-Match header, then the server uses the updated entry's
gd:etag attribute value as an implied
To override the versioning system and update the entry regardless of whether someone else has updated it since you retrieved it, use
If-Match: * instead of specifying the ETag in the header.
For information about which services support strong ETags, see the Migration Guide.
Deleting entries that use strong ETags works much like updating them.
To delete an entry that has a strong ETag, first you retrieve the entry you want to delete, then you send a
DELETE request to the entry's edit URL.
If you want to make sure that you don't delete an entry that has been changed by another client since you retrieved it, then include an HTTP
If-Match header that contains the original entry's ETag value.
If you want to override the versioning system and delete the entry regardless of whether someone else has updated it since you retrieved it, use
If-Match: * instead of specifying the ETag in the header.
If an entry does not have a strong ETag, then a
DELETE request always succeeds.
Partial response (Experimental)
By default, the server sends back the full representation of the target resource after processing requests. Partial response lets you request only the elements or attributes of interest, instead of the full resource representation. This lets your client application avoid transferring, parsing, and storing unneeded fields, so it can use network, CPU, and memory resources more efficiently.
To find out if partial response is available for the product you are using, see its API documentation.
To request a partial response, use the
fields query parameter to specify the elements or attributes you want returned. Here's an example:
The server's response contains only only link and entry elements for the feed; the entry elements contain only ETag, ID, updated, and edit link information. The
fields query parameter syntax is covered in the following sections. For more details about the response, see Handling partial responses.
Note: You can use the
fields query parameter with any request that returns data. In addition to
GET, this includes
PUT (as well as
PATCH, which is used for making partial updates). However, the
fields query parameter only affects the response data; it does not affect the data that you must provide or which fields are updated or created.
Fields parameter syntax summary
The format of the
fields query parameter value is based on XPath syntax; however, it supports only a subset of valid XPath expressions. The supported syntax is summarized below, and additional examples are provided in the following section.
- Use a comma-separated list to select several fields.
a/bto select an element
bthat is nested within element
a/b/cto select an element
'@'prefix to identify an attribute with the given name; omit the
'@'prefix to refer to an element.
- Apply field conditions to select elements that match certain criteria, by placing expressions in brackets "
[ ]" after the element you want to restrict.
fields=entry[author/name='Elizabeth']returns only feed entries for which Elizabeth is the author.
- Specify field sub-selectors to request only specific attributes or sub-elements, by placing expressions in parentheses "
( )" after any selected element.
fields=entry(id,author/email)returns only the ID and the author's email for each feed entry.
- You can delimit strings using either double or single quotes
To escape a double or single quote, repeat the quote
. For example,
"""Hello,"" he said"produces the string
"Hello," he said, and
'''Hello,'' he said'produces the string
'Hello,' he said.
- You can use wildcards in field selections.
entry/gd:*selects all child elements of entry in the
entry/@gd:*selects child element attributes in the same namespace.
fields query parameter acts as an output filter. This means that the partial response is computed only after processing the rest of the query. For example, if you also specify a
max-results query parameter to indicate that you want 20 results per page, then the first 20 results are generated and the partial response is computed from that. If the
fields specification doesn't match any of the first 20 entries selected by the query, then you get back an empty feed; you do not get back the first 20 matching entries.
Note: Do not attempt to use field conditions as query selectors. That is, do not attempt to retrieve a full feed and apply field conditions to filter out items of interest from a very large data set. Whenever possible, use other query parameters, such as
max-results, to reduce the results of each query to a manageable size. Otherwise, the performance gains possible with partial response could be outweighed by the serious performance degradation caused by improper use.
Formatting the fields parameter value
The following guidelines explain how to construct the
fields query parameter value. Each guideline includes examples and provides descriptions of how the parameter value affects the response.
Note: As with all query parameter values, the
fields parameter value must be URL encoded. For better readability, the examples below omit the encoding.
- Identify the fields you want returned, or make field selections.
fieldsquery parameter value is a comma-separated list of elements or attributes (collectively called fields), and each field is specified relative to the root element of the resource representation. Thus, if you are retrieving a feed, fields are specified relative to the
<feed>element, and if you are retrieving a single entry, fields are specified relative to the
<entry>element. If the element you select is (or is part of) a repeating element in the feed, the service returns all instances of that element.
Here are some feed-level examples:
<entry>elements and all sub-elements of those entries, but not any other child elements of
Returns both the feed
<title>element for all feed entries.
Whenever a nested element is returned, the response includes enclosing tags for any
parent elements. The parent tags do not include any other child elements or attributes unless they are also selected explicitly.
Returns only the
<uri>sub-element of the
<author>element for all feed entries.
Returns only sub-elements with local name
ratingin any namespace for all feed entries.
Here are some entry-level examples:
<author>child element of the target entry.
etagattribute of the target entry.
<uri>sub-element of the
<author>element for the target entry.
Returns all sub-fields of
medianamespace for the target entry.
- Restrict the response to selected fields that match certain criteria, or use field conditions.
- By default, if your request specifies an element that occurs more than once, the partial response will include all instances of that element. However, you can also specify that the response should only include elements that have a particular attribute value or elements that fulfill some other condition using "
[ ]" syntax, as shown in the examples below. See the field condition syntax section for more details.
Returns any feed entries that contain a
<link>element with a
relattribute value of
<title>elements that occur in feed entries if their content is
<author>elements that occur in feed entries if they have a
<name>sub-element with content
<author>element in the target entry if it has a
<name>sub-element with content
- Request only parts of the selected elements, or use field sub-selections.
- By default, if your request specifies particular elements, the service returns the elements in their entirety. You can specify that the response should include only certain sub-elements within the selected elements. You do this using "
( )" sub-selection syntax, as in the examples below.
Returns only the
<uri>sub-element for authors in feed entries.
Returns only the
<author>for any entries that have an author name of
Returns only the values of the
hrefattributes for each
<link>element in feed entries.
<link>elements with edit
relattributes for each feed entry.
<title>elements and author
<uri>elements for each feed entry.
More about field condition syntax
You can use field conditions with fields or sub-fields. The condition must evaluate to true for the selected field to be included in the results. If there is no field condition, then all fields of the selected type are included.
The text value of the selected field is used for comparisons. In this context, if the field is an element, the text value is its contents; if the field is an attribute, the text value is the attribute's value. If the field has no text value, then the comparison fails and the field is not included in the results.
The following table shows the XPath operators that are supported for field conditions and provides some examples.
|Date comparison||Use numerical comparison operators, as shown in examples.||
To do date or date/time comparisons, you can cast elements, attributes, or string literals into
Use name of element or attribute as shown in examples.
Booleans can be useful during testing to force field conditions into a true or false state.
Handling partial responses
After a server that supports partial response processes a valid request that includes the
fields query parameter, it sends back an HTTP
200 OK status code along with the requested attributes or elements. If the
fields query parameter has an error or is otherwise invalid, the server returns an HTTP
400 Bad Request status code.
The root element of the response is either
<entry>, depending on the target URI. The root element's content includes only the selected fields for that feed or entry, along with the enclosing tags for any parent elements.
The value of the the request's
fields query parameter can be echoed back in two ways:
- The root element has a
gd:fieldsattribute that shows value of the
fieldsquery parameter specified in the request.
- If the target URI is a feed, each editable entry has a
gd:fieldsattribute that shows the portion of the
fieldsselection that applies to it.
Note: In order to see these
gd:fields attribute values in your partial response, you must include them in your
fields query parameter specification. You can do this explicitly, using
@gd:fields, or using the more general
@gd:*, which also includes ETag information.
The following example query asks the server to return a document that contains only attributes in the
gd namespace (at both the feed and entry level), as well as the feed ID, the title, and the edit link for each feed entry:
The server returns the following partial response, along with a
200 Successful HTTP status code:
<?xml version='1.0' encoding='utf-8'?> <feed xmlns='http://www.w3.org/2005/Atom' xmlns:gd='http://schemas.google.com/g/2005' gd:etag='W/"DEAEQH47eCp7IWA9WxBVGUo."' gd:fields='@gd:*,id,entry(@gd:*,title,link[@rel='edit'])> <id>http://example.com/myFeed</id> <entry gd:etag='"EksPTg1Bfyp7IWA6WhJT"' gd:fields="@gd:*,title,link[@rel='edit']"> <link rel='edit' href='http://example.com/myFeed/1/'/> <title>This year</title> </entry> <entry gd:etag='"EksPQA1Cdyp7IWA6WhJT"' gd:fields="@gd:*,title,link[@rel='edit']"> <link rel='edit' href='http://example.com/myFeed/2/'/> <title>Last year</title> </entry> <entry d:etag='"EksPQAxHeCp7IWA6WhJT"' gd:fields="@gd:*,title,link[@rel='edit']"> <link rel='edit' href='http://example.com/myFeed/3/'/> <title>Today</title> </entry> </feed>
If the selected fields do not match anything, the service still returns a
200 Successful HTTP status code, but the partial response is an empty feed:
<?xml version='1.0' encoding='utf-8'?> <feed xmlns='http://www.w3.org/2005/Atom' xmlns:gd='http://schemas.google.com/g/2005' gd:etag='W/"DEAEQH47eCp7IWA9WxBVGUo."' gd:fields='@gd:*,id,entry(@gd:*,title,link[@rel='edit'])> </feed>
Partial update (Experimental)
Google products that support partial response and editable resources also allow you to use partial update. With partial update, you send only the fields you want to update, rather sending a modified version of the full resource representation. This lets your client application be more efficient when making updates, as well as when using partial response to retrieve data.
Instead of using
PUT, however, you must use a
PATCH request when making a partial update. The semantics for
PATCH are powerful enough to let you add, replace, and delete specific fields for a particular entry, all with a single request.
To find out if partial update is available for the product you are using, refer to the product-specific documentation.
Submitting a partial update request
To submit a partial update request, you send an HTTP
PATCH request to the same URL that you would normally use with
PUT to update the resource. The body of the
PATCH request is a partial
<entry> element that specifies the fields you want to add or modify. The entry's
gd:fields attribute indicates the fields you want to delete.
The server processes
PATCH requests in a specific order:
- It first removes from the resource representation the fields specified by the
The syntax for the
gd:fieldsattribute is the same as for the
fieldsquery parameter used when requesting a partial response. See Supported syntax for more details.
- It then merges into the existing resource representation the data provided in the body of the request.
More details on how the data is merged are provided in Adding or updating fields below.
Note: Since the body of a
PATCH request is not typically compliant with the Atom Syndication Format, the
Content-Type you use with a
PATCH request is
Here is an example of a partial update request:
PATCH /myFeed/1/1/ Content-Type: application/xml <entry xmlns='http://www.w3.org/2005/Atom' xmlns:gd='http://schemas.google.com/g/2005' gd:fields='description'> <title>New title</title> </entry>
PATCH request makes the following changes to the resource representation stored on the server for the target URI's entry:
- Removes the
- Updates the
Semantics of a partial update request
The instructions below explain how to set up your
PATCH request to delete, add, or update specific fields within an entry. A single
PATCH request can perform any combination of these operations.
Deleting fields. Use the
gd:fieldsattribute to identify any fields you want deleted from the resource. The following sample request deletes the title and summary associated with an entry. However the request does not add or update any other data for the entry.
PATCH /myfeed/1/1/ Content-Type: application/xml <entry xmlns='http://www.w3.org/2005/Atom' xmlns:gd='http://schemas.google.com/g/2005' gd:fields='title,summary'/>
Adding or updating fields. Use the body of the
<entry>element to specify the data that you want to add or update for a resource. These fields are merged into the existing data for the resource, after any deletions are made, according to the following rules:
Fields not already present are added. If the resource data does not already specify a value for a field, then the field is added to the existing data. For example, if an entry does not have a title, and your
PATCHrequest contains a
<title>element, then the new title is added to the entry.
Fields already present are replaced or appended. The specific behavior for merging fields that are already specified in the resource data depends on the characteristics of the field:
Non-repeating fields are replaced. If the resource data already specifies a value for a non-repeating element, then the value you specify in the
PATCHrequest replaces the existing value for that element. For example, in the example below, the new title replaces the existing title.
PATCH /myFeed/1/1/ Content-Type: application/xml <entry xmlns='http://www.w3.org/2005/Atom' xmlns:gd='http://schemas.google.com/g/2005'> <title>New Title</title> </entry>
A more complex example is given below. For this example, assume that the entry can have only one author, and that the target resource already has values for the author's name and email address. Even though
<author>element has two child fields, only the
<name>element is present in the data provided. As a result, only that field's value is overwritten. The value of the
<email>element, which is missing from the data provided, remains unchanged.
PATCH /myfeed/1/1/ Content-Type: application/xml <entry xmlns='http://www.w3.org/2005/Atom' xmlns:gd='http://schemas.google.com/g/2005'> <author> <name>New Name</name> </author> </entry>
Repeating fields are appended. If the resource data already specifies a value for a repeating element, then the new element you provide is added to the existing set of values.
Note that there might be times when you want to do something other than add a new instance of a repeating element. For example, you might want to do one of the following:
Replace an entire list of repeating elements. You can delete all the repeating fields using the
gd:fields='ns:accessControl', for example) and provide a complete set of the replacement fields. Since all the existing elements are deleted first, the set of fields you provide do not conflict with any existing values when they are appended.
Replace one value in a set of existing values for a repeating element. In this case, simply remove the single element by defining the
gd:fieldsvalue narrowly enough to avoid deleting other values that you want to keep. For example, to remove only an access control with an
embed, you might use
gd:fields='ns:accessControl[@action="embed"]'. Then you provide the single field that you want to replace it with in the body of the
PATCH /myfeed/1/1/ Content-Type: application/xml <entry xmlns='http://www.w3.org/2005/Atom' xmlns:gd='http://schemas.google.com/g/2005' gd:fields='ns:accessControl[@action="embed"]> <ns:accessControl action="embed" permission="allowed" /> </entry>
Handling the response to a partial update
After processing a valid partial update request, the API returns a
200 OK HTTP response code. By default, the body of the response is the complete entry that you updated. The server updates ETag values when it successfully processes a
PATCH request, just as it does with
PATCH request results in a new resource state that is syntactically or semantically invalid, the server returns an
HTTP 400 Bad Request or
422 Unprocessable Entity HTTP status code, and the resource state remains unchanged. For example, if you attempt to delete a required field and do not provide a replacement, the server returns an error.
Note: It is important to understand how different fields relate to each other. It might be possible to put a resource into an inconsistent state by updating only part of mutually interdependent values. For example, it might be possible to update a start time to a later value than an end time. Although the API should return an error code, we recommend that you fully test these kinds of conditions to ensure consistency.
Alternate notation when PATCH is not supported
If your firewall does not allow
PATCH, then do an HTTP
POST request and set the override header to
PATCH, as shown below:
POST /myfeed/1/1/ X-HTTP-Method-Override: PATCH Content-Type: application/xml ...
Using partial response with partial update
You can use a partial response as the basis of a subsequent partial update request. If you do this, specify a
fields query parameter that includes edit links, as well as
@gd:*. This ensures that the partial response includes information like the ETag and
gd:fields attribute values, which are important for subsequent requests.
Here is an example that returns a partial response that you could use as the basis for a future partial update:
The server responds:
<?xml version='1.0' encoding='utf-8'?> <entry xmlns='http://www.w3.org/2005/Atom' xmlns:gd='http://schemas.google.com/g/2005' gd:etag='"E0UKRAREeCp7IWA6WhJT"' gd:fields="@gd;*,link[@rel='edit'](@href),gd:who"> <link href='http://example.com/myFeed/1/1/'/> <gd:who firstname.lastname@example.org'/> <gd:who email@example.com'/> <gd:who firstname.lastname@example.org'/> </entry>
Suppose that you want to remove the user with email
'email@example.com', add a user with email
'firstname.lastname@example.org', and change the email for the user currently listed as
You can make these changes simply by starting with the results of the previous response, modifying only the fields that are different, and submitting the modified partial entry as the body of the
PATCH request. For this example, the needed modifications are:
<gd:who email='jane'/>from the list of elements provided.
<gd:who email@example.com'/>to the list of elements provided.
PATCH request based on the pevious partial response is shown below:
PATCH /myFeed/1/1/ Content-Type: application/xml <entry gd:fields="@gd:*,link[@rel='edit'](@href),gd:who" gd:etag="FE8LQQJJeSp7IWA6WhVa"> <link href='http://example.com/myFeed/1/1'/> <gd:who firstname.lastname@example.org'/> <gd:who email@example.com'/> <gd:who firstname.lastname@example.org'/> </entry>
Note: This approach relies on the
gd:etag (if supported) attributes being included in the partial response for the entry. The body of the partial entry must retain all fields and attribute that were present in the partial response — except for those you explicitly want to remove. You can update any of the existing fields in the body with new values, and you can include any new fields you want to add.
When a client tries to access a service, it may need to provide the user's credentials to the service, to demonstrate that the user has the authority to perform the action in question.
The approach that a client should use for authentication depends on the type of client:
- A desktop application should use a Google-specific authentication system called Account Authentication for Installed Applications (also known as "ClientLogin"). (Web-based clients should not use this system.)
- A web-based client, such as a third-party front end to a Google service, should use a Google-specific authentication system called Account Authentication Proxy for Web-Based Applications (also known as "AuthSub").
In the ClientLogin system, the desktop client asks the user for their credentials, and then sends those credentials to the Google authentication system.
If authentication succeeds, then the authentication system returns a token that the client subsequently uses (in an HTTP Authorization header) when it sends Data API requests.
If authentication fails, then the server returns a 403 Forbidden status code, along with a WWW-Authenticate header containing a challenge applicable to the authentication.
The AuthSub system works similarly, except that instead of asking the user for their credentials, it connects the user to a Google service that requests credentials. The service then returns a token that the web application can use; the advantage of this approach is that Google (rather than the web front end) securely handles and stores the user's credentials.
Many business logic implementations require session stickiness—keeping track of the state of a user's session.
Google tracks session state in two ways: using cookies, and using a token that can be sent as a query parameter. Both methods achieve the same effect. We recommend that clients support one of these session-state tracking methods (either one is sufficient). If a client doesn't support either of these methods, then that client will still work with Data APIs, but performance may suffer compared to clients that do support these methods. Specifically, if a client doesn't support these methods, then every request results in a redirect, and therefore every request (and any associated data) is sent to the server twice, which affects the performance of both the client and the server.
The Google client libraries handle session state for you, so if you use our libraries, you don't have to do anything to get session state support.
You may find the following third-party documents useful:
- Comparison of Atom and RSS from intertwingly
- Overview of Atom from IBM
- Dublin Core extensions to RSS
- HTTP 1.1 method definitions; specification for
- HTTP 1.1 status code definitions
- How to Create a REST Protocol
- Building Web Services the REST Way
- A Technical Introduction to XML
- XML Namespaces by Example
- ETag section of HTTP specification