This document introduces basic search concepts, and describes
advanced techniques that produce more efficient search results.
Topics:
Getting
Started with Search
Spelling
Capitalization
Common
Words
Date
Sort
Numbers
Widening Your
Search
Refining Your
Search
Word
Exclusion
Phrase
Searches
Range
Searches
Number
Ranges
Date
Ranges
Metadata and
Meta Tag Ranges
Advanced
Search Operators
Getting Started
with Search
To search for a document, type a few descriptive words in the search box,
and press the Enter key or click the search button. A
results page appears with a list of documents and web pages that are
related to your search terms, with the most relevant search results
appearing at the top of the page. By default, only pages that include
all of your search terms are returned. So to broaden or restrict the
search, include fewer or more terms. You do not need to include "and"
between the terms. For example, to search for engineering product
specification documents, type the following:
The search appliance uses sophisticated text-matching techniques to find
pages that are both important and relevant to your search. For instance,
the search appliance analyzes not only the candidate page, but also the
pages that link to it, too. The search appliance also prefers pages in
which your query terms are near each other. Every search result lists
one or more snippets, or excerpts from the document, to display the
search terms in context. In the snippet, your search terms are
displayed in bold text so that you can quickly determine if that result
is from a page or document you want to visit.
Note: Encrypted, viewable PDF documents are converted to HTML for
indexing, but the HTML is not displayed.
Spelling
For searches in some languages, a single spelling suggestion is returned
with the results for queries where the spell checker has detected a
possible spelling mistake. The spell checker supports the following
languages by default:
U.S. English
French
German
Italian
Brazilian Portuguese
Spanish
The spell checker feature is context sensitive.
Your browser's language setting affects how the search appliance handles
spelling corrections.
If the browser language is English, Portuguese, French, Italian,
German, or Spanish, the search appliance returns spelling corrections.
If the browser language is Japanese, traditional Chinese, Korean, or
simplified Chinese, the search appliance returns English spelling
corrections.
If the browser language is any other language, the search appliance
does not return spelling corrections.
For information about how to change your browser's language setting, read
the help system for the browser.
Capitalization
The search appliance searches are not case sensitive. All
letters, regardless of how you enter them, are handled as lower case. For
example, searches for "george washington," "George Washington," and
"George washington" return the same results.
Common Words
Because they tend to slow down your search without improving
the results, the search appliance ignores some terms, including:
Common words and characters, such as "where" and "how," when they are
used in conjunction with other search terms
For example, if you search for "who," the search appliance does not
ignore it. The search appliance returns results for "who." However, if you
search for "Pat who," the search appliance does ignore "who" and only
returns results for "Pat."
Certain single digits and single letters
If a common word is essential to getting the results you want, you can
include it by putting a plus ("+") sign in front of it. Include a space
before the "+" sign, but not after it. For example, to search for documents
about Star Wars I, type the following:
Alternatively, you can enclose a series of words with quotation marks and
do a phrase
search.
Date Sort
By default, search results are sorted by relevance, with the most
relevant result appearing at the top of the page. If you want to sort the
documents by date instead, click the Sort by Date link. The most recent
document appears at the top of the page and the date of each file is
returned in the results. Results that do not contain dates are
displayed at the end and are sorted by relevance.
Numbers
When you search for numbers, do not use exponential numbers, such as
"1e10," or negative integers, such as "-12."
Numbers that are separated by commas are treated as separate figures, not
fractional numbers; that is, the comma is treated as a term separator, not
a decimal separator. For example, if you type "3,75", the search query is
treated as a search for two separate terms, "3" and "75", not the decimal
fraction, "three and three quarters." Commas that separate every three
digits are ignored and are not necessary. For example, both "10,000" and
"10000" are treated alike.
Widening Your
Search
You can expand your search by using the OR operator. To retrieve
pages that include either word A or word B, use an uppercase OR between terms. For example, to search for an office in either London or
Paris, type the following:
Refining Your
Search
Since the search appliance returns only web pages that contain all
of the words in your query, refining or narrowing your search is as simple
as adding more words to the search terms you have already entered. The
refined query returns a subset of the pages that were returned by your
original broad query. If that does not get the results that you want, you
can try to exclude words, search for exact phrases, or restrict the search
to a range of numbers. These techniques are described in the following
subsections.
Word Exclusion
If your search term has more than one meaning, you can focus your
search by adding a minus sign ("-") in front of words related to the meaning
you want to avoid. Make sure you include a space before the minus sign.
You can daisy chain a list of words you want to exclude.
For example, to search for the planet Saturn and exclude search results
about the car company or Roman god, type the following query:
The search appliance returns pages about Saturn that do not contain the
word "car" or "god."
Phrase
Searches
Phrase searches are useful when you are searching for famous sayings or
specific names. You can search for an exact phrase or name in the following
ways:
By enclosing the phrase in quotation marks. The search appliance only
returns documents that include the exact phrase you entered.
By using phrase connectors—such as hyphens, slashes, periods, equal
signs, and apostrophes—in between every word of your
search query.
Phrase connectors and quotation marks join your search words as a single
unit. For example, if you type the following query, the search
appliance treats it as a phrase search even though the search
words are not enclosed in quotation marks.
Range Searches
You can confine your search query within a certain range. You can set
ranges for dates, weights, prices, meta tags, and so on. The following
subsections describe ways you can refine your searches with
ranges.
Number Ranges
To search for documents or items that contain numbers within a range,
type your search term and the range of numbers separated by two periods
(".."). You can set ranges for weights ("250..500 g carbon fork"),
dimensions ("90..100 mm stem"), years ("tour de france 2000..2006"),
prices in dollar currencies only ("bike lights $10..$30"), and so on. Be
sure to specify a unit of measurement or some other indicator of what the
number range represents.
For example, to search for pencils that costs between $1.50 and
$2.50, type the following:
Each number in the range should not include more than six significant
digits. For example, if you were to type the search query, "1..1234567 ton
truck," only the first six significant digits in the "1234567" would be
included in the range search; that is, it is as though you have just
typed, "1..1234560 ton truck."
Date Ranges
You can search for documents that contain dates that fall within a time
frame. To use date range search, type all of the following:
The search term
The daterange: operator
The start date
The range separator (which is two periods if you are using a
YYYY-MM-DD format or a hyphen if you are using a Julian format)
The end date
Do not add a space between the search operator and the date range. The
dates could be in either of the following formats:
The YYYY-MM-DD (ISO 8601) format. Date ranges using this format
should be separated by two periods ("..").
The Julian format. The Julian date is calculated by the number of
days since January 1, 4713 BC. For example, the Julian date for August
1, 2001 is 2452122. Date ranges in this format should be separated by a
hyphen ("-").
For example, to search for a document about Harry Potter that was
modified within a specific two-year period, type the following:
The earliest date that you can use in your date range search is January
1, 1900; and the latest date, June 6, 2079.
Metadata and Meta Tag Ranges
You can search only for documents that include metadata or meta tags
that contain numbers within the range you specified. To use metadata range
search, type all of the following:
The search term
The inmeta: operator
The name of metadata or meta tag
The range of numbers separated by two periods ("..")
For accurate date range searches with inmeta, the meta tag
content must contain only the date and no other data. Suppose your
documents have metadata called "modified" that contains the last modified
dates of the documents. To search for a document about risks that was
created sometime in 2006, you could type the following:
You can use the inmeta operator beyond just searching for
documents with metadata that includes a range of dates or numbers. To
learn more about inmeta, see the Search Protocol Reference on
the public
search appliance documentation page.
Advanced Search
Operators
The search appliance supports several advanced operators, which are query
words that restricts your search to a smaller set of documents. When you
enter your search query, do not add a space between the search operator
and the search terms.
Search Operator
Description
Example
allinanchor:
Restricts the search to pages that contain all the search terms in
the anchor text of the page. The following example
shows an anchor tag:
<a
href="http://foo.com">Go Foo</a>
allinanchor: evaluates the text between > and </a>. allinanchor: evaluates only <a href anchor tags. It does not evaluate <a name anchor tags.
An anchor is a marker inserted at a specific section of a page. It
lets the writer of the document create links to these anchors, which
quickly take the reader to the specified section. The table of contents
at the top of this document, for example, uses hyperlinks to anchors
embedded throughout this document.
Do not include any other search operators with the allinanchor: operator.
Typing allinanchor:cheap books in the search box returns
only pages that have anchor text that include the words "cheap" and
"books" between > and </a>.
allintitle:
Restricts the search to documents whose HTML title contains all the
search terms.
Typing allintitle:google search in the search box returns
only documents that have both "google" and "search" in the HTML
title.
allintext:
Restricts the search to documents whose titles or body text contains
the search terms.
Typing allintext:google search in the search box returns only documents that
have both "google" and "search" in the title or body text of the
document.
allinurl:
Restricts the search to documents whose URL contains the search
terms. The search operator does not require the query words to be
adjacent to each other in the document, nor does it require the words to
appear in a particular order in the document.
The search operator works on words in the URL, not URL components
such as punctuation. Slashes ("/"), for example, are ignored.
Typing allinurl:google search in the search box returns only
documents that have both "google" and "search" in the URL.
Typing allinurl:google/search in the search box returns the
same documents as the previous example. The slash in the search term is
altogether ignored.
cache:
The search engine keeps the text of the many documents it crawls
available in "cache." A cached version of a web page can be retrieved if
the original page is unavailable, such as when the page's server is
down. The cached page appears exactly as it looked when the crawler last
crawled it, but it includes a message (at the top of the page) to
indicate that it's a cached version of the page.
If you include search words in addition to the web address in your
query, those search words will be highlighted within the cached
document.
Typing cache:www.google.com in the search box returns the
cached version of Google's homepage.
Typing cache:www.google.com press releases in the search box
returns the cached content with the words "press" and "releases"
highlighted.
daterange:
Restrict search to documents that contain dates that fall within a
time frame, or before or after a specified date. You can search any
dates from 1900-01-01 to 2079-06-06. The dates can be in any format
except Julian format, which is considered a number range search. When
specifying dates in ISO 8601 format (YYYY-MM-DD), separate date ranges
by two periods (..).
Typing election daterange:2004-01-13..2006-01-13in the search box returns results for the search term "election"
whose dates fall within the specified time range.
filetype:
Restricts the search to specific file types such as Excel
spreadsheets, PDF files, or Word documents. Type the filetype: operator followed by the file extension.
Typing cars filetype:pdf in the search box returns only PDF
files about cars.
info:
Returns the following information for that particular URL:
The cache of the page
Web sites that are similar to the page
Web pages that have hyperlinks to the page
Web pages that are hyperlinked in that page
Web pages that contain the URL in its body text
Typing info:www.google.com in the search box returns the
following information about the Google home page:
The cache of www.google.com
Web sites that are similar to www.google.com
Web pages that have hyperlinks to www.google.com
Web pages that are hyperlinked in www.google.com
Web pages that contain the www.google.com in its body
text
inmeta:
You can filter results by meta tags and their values using inmeta. Used with the operators ~ or =, inmeta restricts results to required or
partial meta tag values in the same way as the requiredfields and partialfields search
parameters.
Typing inmeta:department=Human Resourcesreturns documents where the meta tag includes department=Human Resources.
intext:
Restricts the search to documents that contain the search word in the
titles or body text of the documents. The search appliance does not
search for the query word in the metadata,
anchors, or urls.
Putting intext: in front of every word in your query is
equivalent to putting allintext: at the front of your query.
Typing intext:google returns documents that mention the word "google"
in their title or body text.
Typing intext:google intext:search in the search box is the
same as typing allintext: google search.
intitle:
Restricts the search to documents that contain the search word in the
HTML title.
Putting intitle: in front of every word in your query is
equivalent to putting allintitle: at the front of your query.
Typing intitle:google search returns documents that mention
the word "google" in their HTML title, and mention the word "search" in
the title, body text, anchor, or anywhere else in the document.
Typing intitle:google intitle:search in the search box is
the same as typing allintitle:google search.
inurl:
Restricts the search to documents that contain the search word in the
URL.This operator works on words, not URL components such as
punctuation. Slashes ("/"), for example, are ignored.
Putting the inurl: operator in front of every word in your
query is equivalent to putting allinurl: at the front of your query.
Typing inurl:google search in the search box returns
documents that mention the word "google" in their URL and mention the
word "search" in the URL, body text, title, or anywhere else in the
document.
Typing inurl:google/search in the search box returns the
same documents as the previous example. The slash in the search term is
altogether ignored.
Typing google inurl:google inurl:search in the search box
returns documents that contain both "google" and "search" in the URL. It
returns the same documents as the search query allinurl:google
search.
link:
Restricts the search to all pages that link to the web site in the
query.
No other search term can be appended to this search operator and the
specified web site.
Typing link:www.berkeley.edu in the search box returns all
the pages that link to that page.
site:
Restricts the search to documents in a web site. If you do not
specify the web site and just type the generic top-level domain, such as
.com, .edu, or .org, the search engine returns all documents in the
generic top-level domain.
The site: operator lets you extend the search restriction
down to directories.
Typing help site:www.google.com in the search box returns
pages about help or user documentation within www.google.com.
Typing help site:com in the search box finds pages about
help or user documentation within all web sites that end in .com.
Typing site:www.google.com/enterprise/ restricts the search
to everything at the enterprise directory level. If the trailing slash
is omitted, as in www.google.com/enterprise, all subdirectories
are searched.
wildcard:
The wildcard: operator enables you to search by word pattern rather
than the exact spelling of a term. The search appliance supports two
wildcard operators:
*--Matches zero or more characters
?--Matches exactly 1 character
A wildcard query term must satisfy at least one of the following
conditions:
A sequence of at least 2 characters at the start of a word
A sequence of at least 2 characters at the end of a word
A sequence of at least 3 characters anywhere in the
word
Typing wildcard:go* matches any words that begin with the
letters "go".
Typing wildcard:g? matches any single character that follows
the letter "g".
Typing wildcard:*le matches any words that end with the
letters "le".
Typing wildcard:*ear* matches any words that contain the
letter sequence "ear" anywhere in the word.
