Thank you Kristina,

I know about the URL structure, I have been trying the past few months to cleanse this site that I was not involved in its creation.  It has several more SEO problems that have either been fixed or not yet, but we are talking about more than 50 SEO problems I've found so far - most of these critical.

On the sitemap that I built, the junk pages do not exist, and because this is sitemap I have written myself, I can easily make another containing the articles that I have removed (just reverse a part of my select query for the sitemap to get the ones I have removed).

http://www.neakriti.gr/webservices/sitemap-index.aspx

So far I implemented the last of your suggestions and here is an example:

This is a valid article page
http://www.neakriti.gr/?page=newsdetail&DocID=1314221 - (Status Code: 200)

This is a non existent article page (never existed at the first place) - (Status Code: 404)
http://www.neakriti.gr/?page=newsdetail&DocID=12345678

This is one of the articles that I removed from sitemap and site - (Status Code: 410)
http://www.neakriti.gr/?page=newsdetail&DocID=894052

Also I would like you to take a look at another question about the same site and see that it can relate to this question with garbage articles too...
https://moz.com/community/q/multiple-instances-of-the-same-article

Thank you so much!

Thank you,

so you suggest that based on my date based query, instead of blocking everything before that date blindly, keep blocking it with 410, while anything that doesn't exist anyway return 404.

Also another question, about the blocked articles that return 410, should I put their URLs back on the xml sitemap or not?

Thank you for your response.

I defenintelly cannot use noindex because as I explained I changed all articles prior to the minimum given date to return 404.  So this content is not visibly available on the web in order to contain a noindex directive.  Unless you mean to have it at my custom 404 page, where yes its there.

Also there is no folder to associate in robots, since they are in ugly form of GET params like DOCID=12345.  So given that, there are thousands of DocIDs that are junk and removed, and thousands that are the actuall articles.

So I assumed that creating a "deleted articles" sitemap where each <url>will contain an <expires>2016-06-01</expires> tag seemed the most logical thing, but I am afraid its for "custom search engines", rather than for normal de-index requests as its provided bellow</url>

https://developers.google.com/custom-search/docs/indexing#on-demand-indexing

Hi,

I am trying to cleanse a news website.  When this website was first made, the people that set it up copied all kinds of articles they had as a newspaper, including tests, internal communication, and drafts.  This site has lots of junk, but this kind of junk was on the initial backup, aka before 1st-June-2012.  So, removing all mixed content prior to that date, we can have pure articles starting June 1st, 2012!

Therefore

1. My dynamic sitemap now contains only articles with release date between 1st-June-2012 and now
2. Any article that has release date prior to 1st-June-2012 returns a custom 404 page with "noindex" metatag, instead of the actual content of the article.

The question is how I can remove from the google index all this junk as fast as possible that is not on the site anymore, but still appears in google results?

I know that for individual URLs I need to request removal from this link
https://www.google.com/webmasters/tools/removals

The problem is doing this in bulk, as there are tens of thousands of URLs I want to remove.  Should I put the articles back to the sitemap so the search engines crawl the sitemap and see all the 404?  I believe this is very wrong.  As far as I know this will cause problems because search engines will try to access non existent content that is declared as existent by the sitemap, and return errors on the webmasters tools.

Should I submit a DELETED ITEMS SITEMAP using the <expires>tag? I think this is for custom search engines only, and not for the generic google search engine.
https://developers.google.com/custom-search/docs/indexing#on-demand-indexing</expires>

The site unfortunatelly doesn't use any kind of "folder" hierarchy in its URLs, but instead the ugly GET params, and a kind of folder based pattern is impossible since all articles (removed junk and actual articles) are of the form:
http://www.example.com/docid=123456

So, how can I bulk remove from the google index all the junk... relatively fast?