Well, no, because networked software. In addition, due to the FUCKING EXORBITANT cost of most modern ILS systems many libraries (public, private, and otherwise) enter into consortiums whereby they share one central ILS system among the group and compartmentalize the data such that one library's patrons only see the items held by that library and not items held by the other consortium members.
Bibliographic data in libraries is for the most part kept in the MARC format
(which is old and creaky but whatevs) and for every item in the database a record like the following is available for whatever data manipulation you care to whip it with:
=LDR 02390cam 2200433Ia 4500
=020 \$a9781400826773 (electronic bk.)
=020 \$a1400826772 (electronic bk.)
=100 1\$aLavi, Shai Joshua.
=245 14$aThe modern art of dying$h[ebook] :$ba history of euthanasia in the United States /$cShai J. Lavi.
=260 \$aPrinceton, N.J. ;$aWoodstock :$bPrinceton University Press,$c2007.
=300 \$a1 online resource (x, 226 p.) :$btables.
=500 \$aOriginally published: 2005.
=504 \$aIncludes bibliographical references and index.
=520 \$aHow we die reveals much about how we live. In this provocative book, Shai Lavi traces the history of euthanasia in the United States to show how changing attitudes toward death reflect new and troubling ways of experiencing pain, hope, and freedom. Lavi begins with the historical meaning of euthanasia as signifying an "easeful death." Over time, he shows, the term came to mean a death blessed by the grace of God, and later, medical hastening of death. Lavi illustrates these changes with compelling accounts of changes at the deathbed.
=588 \$aDescription based on print version record.
=650 \0$aEuthanasia$zUnited States$xHistory.
=655 \4$aElectronic books.
=856 40$yFULL TEXT ONLINE$uhttp://search.ebscohost.com/login.aspx?direct=true&scope=site&db=nlebk&db=nlabk&AN=286607
So for every item contained in those libraries there is a corresponding record available in the database that can potentially encode all sorts of information such as when it was purchased, when it was removed and by whom, etc. etc. That data is, or should be, readily available to the database admin, and it does not go away once the item itself has been destroyed (natch). Which makes me wonder why the Canadian Government hasn't offered it up (or offered a breakdown of such) to prove their statements of duplicates removed and unique items transferred and no book burning whatsoever.
And if I might go a touch further, as an admin myself, if my boss came to me tomorrow and told me to delete the entire bibliographic database, it would be trivial to dump this information (i.e. a list of every item contained in the library) to a large text file for use or storage elsewhere. My library is a small one (~50K volumes) and the corpus of our DB in the compressed MARC format takes up about 15MB.