Different results between Google API and "real" Google

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Just do a simple search in Google
2. The same search with the Google API
3. Different results between both of them, whatever the request is or the
language is (the hl parameter)

What is the expected output? 
The same first results

What do you see instead?
Totally different results

What version of the product are you using? On what operating system?
Google AJAX Search API 1.0 vs Google. All operating systems.

ex :
http://www.google.fr/search?hl=fr&q=livre&btnG=Recherche+Google&meta=
http://ajax.googleapis.com/ajax/services/search/web?v=1.0&hl=fr&q=livre

I've read the documentation and I don't see any other parameter that would
change the results, so I guess it's an issue, and if it's not an issue then
you should probably mention in the documentation that we're not supposed
the get the same results between the two of them.

Thanks a lot, 

Ugo PARSI

Original issue reported on code.google.com by ugo.pa...@gmail.com on 29 May 2008 at 1:33

[GoogleCodeExporter]

This sure looks like a duplicate of bug 27, which is now fixed.  Can anyone 
verify?

Jeremy R. Geerdes
Effective website design & development
Des Moines, IA

For more information or a project quote:
http://jgeerdes.home.mchsi.com
jgeerdes@mchsi.com

If you're in the Des Moines, IA, area, check out Debra Heights Wesleyan Church!

Original comment by jrgeer...@gmail.com on 3 Jun 2008 at 4:30

[GoogleCodeExporter]

Still doesn't work on my case.

Greetings, 

Ugo PARSI

Original comment by ugo.pa...@gmail.com on 3 Jun 2008 at 5:12

[GoogleCodeExporter]

Oh, of course.  My fault!  This isn't a dup of 27.  I read right past the .fr 
part of
the url.  The API currently only searches www.google.com; i.e., there's no way 
to
restrict it to a localized version (e.g., google.fr).  It is, however, a 
suggestion:

http://code.google.com/p/google-ajax-apis/issues/detail?id=24&colspec=ID%20Type%
20Stars%20Status%20Priority%20Opened%20Owner%20Summary

Could we classify it as a duplicate of that?

jg

Original comment by jrgeer...@gmail.com on 3 Jun 2008 at 5:17

[GoogleCodeExporter]

Hmmm I don't understand, the documentation is talking about a "language" 
parameter
(hl) and if I change it, I've got different results anyway (between french, 
german
and english).
The problem is that I've got different results, no matter what...
If I am asking English, I've got different results from google.com, the same 
for .de
with Deutsch, .fr with French, etc..., etc...

Greetings, 

Ugo PARSI

Original comment by ugo.pa...@gmail.com on 3 Jun 2008 at 5:33

[GoogleCodeExporter]

Right; I understand that.  There are differences between how www.google.com 
and, for
example, www.google.fr rank their searches.  For instance, if you search for
"football" at google.com, you'll find a bunch of things about the National 
[American]
Football League, but if you search for "football" at www.google.co.uk, you'll 
find a
bunch of stuff about soccer, as one would expect.  Even if you restrict to a 
specific
language, those differences persist between local versions, so since the AJAX 
Search
API will always search google.com, you will likely never (or, at the very 
least, very
rarely) see the same results between it and one of the other national domains.

That's what I'm saying.  The suggestion I linked to was that the API be able to
switch to use a particular localized Google (e.g., google.fr).  I'm wondering 
if your
bug stems from its inability to do that.

jg

Original comment by jrgeer...@gmail.com on 3 Jun 2008 at 5:40

[GoogleCodeExporter]

Original comment by jrgeer...@gmail.com on 4 Jun 2008 at 12:44

• Added labels: APIType-Search

[GoogleCodeExporter]

Partial overlap with #4 as well, I think.

Original comment by asheme...@gmail.com on 6 Jun 2008 at 2:40

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

The hl variable shouldn't change a thing about the results themselves, it only
changes the language of the google interface/presentation. It's the lr variable 
which
restricts the search results to a specific language. The lr variable btw also 
depicts
an incorrect amount of search results.

1) no variables set; result amount is consistently too low by a factor of
approximately 10, when the actual result is more then say 5 Billion the API 
actually
gives a reasonable approximation
2) with the language variable set; result amount inconsistently too low, factor
ranges from 1 to 200. Very often the results amount is lower then 1.000.000 and 
in
the range of 300.000 - 400.000 .

No it's not related to localization, using local google version still give
inconsistent difference, seems more like somebody messed up at the db level.

Original comment by brami...@gmail.com on 29 Jul 2008 at 10:55

[GoogleCodeExporter]

@bramiozo, hl does change the search results :
http://ajax.googleapis.com/ajax/services/search/web?v=1.0&hl=en&q=bonjour
gives Apple's bonjour as 1st result
http://ajax.googleapis.com/ajax/services/search/web?v=1.0&hl=es&q=bonjour
gives a wikipedia article

These results are the same in the html version :
http://www.google.com/search?q=bonjour&hl=en 
http://www.google.com/search?q=bonjour&hl=es 

The strange thing is with the french hl, the ajax api gives english results :
http://ajax.googleapis.com/ajax/services/search/web?v=1.0&hl=fr&q=bonjour
(the first result in the html version is "paruvendu.fr" instead of apple.com)

Original comment by tangui.m...@gmail.com on 10 Aug 2008 at 11:01

[GoogleCodeExporter]

For the result count it shouldn't matter what hl value you use, try it with 
native
terms, say Dutch for bicycle "fiets".

I am talking about differences of a factor 1-200 for specific native terms,
no way this is related to localisation.  

This should be resolved or at least explained in more detail by a google staff 
member
because right now the results number is worthless and if I recall it correctly 
this
issue has been floating around for many months.

Original comment by brami...@gmail.com on 30 Aug 2008 at 3:29

[GoogleCodeExporter]

I also face the same issue, not only with AJAX even SOAP apis also have the 
same 
problem. I suspect there is something different processing happening for 
browser 
results (different data sources may be) and api results.. Google not 
acknowledging this fact at all.  

Original comment by hajamaid...@gmail.com on 15 Oct 2008 at 3:17

[GoogleCodeExporter]

I didn't found any parameter to make it site search.

Is there any parameter to do that.

If there why dont they given in documentation?

Original comment by w3.anilk...@gmail.com on 15 Oct 2008 at 11:17

[GoogleCodeExporter]

I didn't found any parameter to make it site search.

Is there any parameter to do that.

If there why dont they given in documentation?

Original comment by w3.anilk...@gmail.com on 15 Oct 2008 at 11:17

[GoogleCodeExporter]

w3.anilkumar-

Is your post related to the rest of this thread, or are you starting a new 
subject? 
At any rate, you want to include site:mysite.com in your query to restrict your
search to a given site.  It's probably not documented in the API documentation
because it's documented in the general search documentation.

jg

Original comment by jrgeer...@gmail.com on 15 Oct 2008 at 12:13

[GoogleCodeExporter]

isnt it possible to get more that 4 results using 
"http://ajax.googleapis.com/ajax/services/search/web?v=1.0&q=lkj" ?

Original comment by marcel.a...@googlemail.com on 21 Nov 2008 at 11:01

[GoogleCodeExporter]

isnt it possible to get more that 4 results using 
"http://ajax.googleapis.com/ajax/services/search/web?v=1.0&q=lkj" ?

Original comment by marcel.a...@googlemail.com on 21 Nov 2008 at 11:09

[GoogleCodeExporter]

I don't know what it has to do with the rest of the issue, but yes, you can get 
8
results by adding "&rsz=large" to the end of that url.

Original comment by jrgeer...@gmail.com on 21 Nov 2008 at 11:14

[GoogleCodeExporter]

Issue 123 has been merged into this issue.

Original comment by lisba...@google.com on 6 Jan 2009 at 9:58

[GoogleCodeExporter]

I am experiencing this problem currently.

Use the following query in both "regular" Google and then with the API:

?
hl=en&q=site:beeradvocate.com%20ebc+blueberry+%22review+this+beer%22+%2Binurl%3A
%22beer%2Fprofil
e%22&rsz=small

To make things easy:

1. Lots of results:

http://www.google.com/search?
hl=en&q=site:beeradvocate.com%20ebc+blueberry+%22review+this+beer%22+%2Binurl%3A
%22beer%2Fprofil
e%22

2. No results:

http://ajax.googleapis.com/ajax/services/search/web?
v=1.0&hl=en&q=site:beeradvocate.com%20ebc+blueberry+%22review+this+beer%22+%2Bin
url%3A%22beer%
2Fprofile%22

Original comment by davega...@gmail.com on 18 Mar 2009 at 9:45

[GoogleCodeExporter]

Also confirming the same issue. 

Search for this book title "REMINISCENCES  INCIDENTS IN THE LIFE OF A PIONEER 
IN OREGON AND IDAHO" in www.google.com or via the 
API and you get completely different results.

http://www.google.com/search?hl=en&q=REMINISCENCES++INCIDENTS+IN+THE+LIFE+OF+A+P
IONEER+IN+OREGON+AND+IDAHO&btnG=Search
1. www.worldcat.org
2. www.pbagalleries.com
3. www.alibris.co.uk
4. www.amazon.ca
etc.

http://ajax.googleapis.com/ajax/services/search/web?
v=1.0&rsz=large&q=REMINISCENCES++INCIDENTS+IN+THE+LIFE+OF+A+PIONEER+IN+OREGON+AN
D+IDAHO&btnG=Search
1. oregonoverland.com
2. www.historicoregoncity.org
3. www.idahohistory.net
4. whmc.umsystem.edu
etc.

Original comment by howb...@gmail.com on 20 Mar 2009 at 3:32

[GoogleCodeExporter]

Is this bug being checked into? 

Original comment by wired...@gmail.com on 23 Mar 2009 at 4:26

[GoogleCodeExporter]

> Is this bug being checked into? 

The bug status is still in "new" (aka "unopened") status, it was created 10 
months 
ago, and a developer still has not been assigned to the bug. So I can pretty 
definitely state that the answer is "no," without being a Google representative 
myself. 

Original comment by howb...@gmail.com on 23 Mar 2009 at 5:22

[GoogleCodeExporter]

Yeah, it's been a while ;)

Original comment by ugo.pa...@gmail.com on 23 Mar 2009 at 5:26

[GoogleCodeExporter]

Issue 223 has been merged into this issue.

Original comment by lisba...@google.com on 8 Apr 2009 at 4:32

[GoogleCodeExporter]

Issue 181 has been merged into this issue.

Original comment by lisba...@google.com on 8 Apr 2009 at 4:33

[GoogleCodeExporter]

Very easy to reproduce:

1) Go to http://www.google.com/uds/samples/apidocs/helloworld.html
2) Enter the following phrase search:

"plague is he, all must keep their distance or fear"

3) Observation: No results.

Try the same on google.com -- you'll see many results.  Note there are
twenty-eight sites with that exact text on it.  Its part of a famous
excerpt from the "The Raven" poem.

Original comment by brian.k...@gmail.com on 8 Apr 2009 at 5:22

[GoogleCodeExporter]

Issue 215 has been merged into this issue.

Original comment by jrgeer...@gmail.com on 15 Apr 2009 at 10:04

[GoogleCodeExporter]

Issue 214 has been merged into this issue.

Original comment by jrgeer...@gmail.com on 15 Apr 2009 at 10:04

[GoogleCodeExporter]

Comment 27 from this issue is very similar to Issue 27 that was apparently 
resolved.

However comment 10 on Issue 27 suggests that there is still an issue.

Original comment by sam.gra...@io-me.com on 22 Apr 2009 at 8:52

[GoogleCodeExporter]

The issue that is NOT resolved is when you are NOT doing a site search. See the 
multiple examples listed above that do not include "site:domain." It takes a 
matter 
of seconds to repeat the test to verify that it's still an open issue.

Original comment by howb...@gmail.com on 23 Apr 2009 at 12:21

[GoogleCodeExporter]

How can we increase the priority of this or at least have a Google developer 
take a
look at it?

Original comment by csgyuri...@gmail.com on 7 May 2009 at 6:12

[GoogleCodeExporter]

If it helps, I can make a keg of beer mysteriously appear outside of building 
42.  No questions asked...

Original comment by brian.k...@gmail.com on 7 May 2009 at 7:03

[GoogleCodeExporter]

Also confirming the same issue. 

Original comment by angelo%r...@gtempaccount.com on 15 May 2009 at 2:32

[GoogleCodeExporter]

Is there anybody at google who reads these threads sometime? Please give us a 
sign. :-)

Original comment by InstantR...@gmail.com on 30 May 2009 at 6:15

[GoogleCodeExporter]

I'm experiencing the same bug. Pretty much any query I do gives different 
results in 
the API search vs. google.com. This <em>is</em> including the site:paratemeter. 

Example:

biochemistry site:http://uwaterloo.ca

Results are more or less the same but order is different. This happens with 
many 
terms (try also "history", "communications"). Usually the results are mostly 
the 
same but the amount of variation differs between search terms.

Original comment by mjjac...@gmail.com on 24 Jun 2009 at 6:43

[GoogleCodeExporter]

Google apparently doesn't listen to their users at all. Such a disappointment.

Original comment by flamefus...@gmail.com on 22 Jul 2009 at 10:47

[GoogleCodeExporter]

This looks the same as my Issue 250 raised on May 18 2009

Original comment by webmas...@cixvfrclub.org.uk on 4 Aug 2009 at 9:47

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

This issue and Issue 78 , combined with the disappointing lack of communication 
from 
the Google does not allow this AJAX Search API to be a viable tool.  If the 
results can 
not be compared with doing a manual google web search (localized or not), then 
what is 
the benefit of using the Google API over say Yahoo 
(http://developer.yahoo.com/search/) 
or Bing (http://www.bing.com/developers) Now, where is the "Tweet This" button?

Original comment by bade...@gmail.com on 11 Aug 2009 at 7:55

[GoogleCodeExporter]

I am having similar situation with site: operator. Google AJAX API results are
different than web results..
I have create a thread here:
http://groups.google.com/group/google-ajax-search-api/browse_thread/thread/71918
5ac5fd86af0/7effd32922d4e55d?hl=en#7effd32922d4e55d

Thanks

Original comment by jajugir...@gmail.com on 31 Aug 2009 at 7:13

[GoogleCodeExporter]

What's the point in this API service if it doesn't return the expected results?
Yahoo! and Microsoft obviously understand the importance - it's a deal-breaker.

Original comment by cjh...@gmail.com on 18 Oct 2009 at 11:41

[GoogleCodeExporter]

Does google change it ot it must work like now?

Original comment by alexzapo...@gmail.com on 19 Oct 2009 at 6:10

[GoogleCodeExporter]

We are facing same issue on Image search.... Any updates from Goolge if it is 
being 
fixed soon?

Bes Regards
Imran Tamboi.

Original comment by itamb...@gmail.com on 19 Oct 2009 at 6:19

[GoogleCodeExporter]

Maybe some one from google developers can say something?

Original comment by alexzapo...@gmail.com on 19 Oct 2009 at 6:29

[GoogleCodeExporter]

They use fancy javascript to stop people from using stuff like CURL to scrape 
results
directly and then make the only remaining method inconsistant at best. Epic 
fail...

Original comment by Lococo...@gmail.com on 30 Oct 2009 at 6:31

[GoogleCodeExporter]

I'm having the same issue. Looking for http://www.walksguide.com
I'm getting different results from UI than API
It's blocking

Original comment by JordiSu...@gmail.com on 5 Nov 2009 at 10:38

[GoogleCodeExporter]

Solve this issue please))!!

Original comment by Romankon...@gmail.com on 6 Nov 2009 at 12:54

[GoogleCodeExporter]

[deleted comment]