I found many of the commercial databases surprisingly flawed. Better to use dblp computer science bibliography.
I remember doing a thorough analysis of papers published at ECRTS for the 25th anniversary edition, looking for most active authors, collaborations and their development etc over time. To my surprise the commercial ones had errors, e.g. IEEE Xplore with typos in author names, affecting publication counts, or Google Scholar breaking titles in various ways, each resulting in a separate publication etc. In the end, we had to hand edit publication lists from the proceedings for correct results.
So the “service” of for profit publishers, for which we have to pay (see post on open access) is not very good, Google and co let us do the work of curating their sloppy work, as usual for free, so they can profit from it.
Instead I recommend the computer science bibliography dblp, the on-line reference for bibliographic information on major computer science publications, now operated and maintained by Schloss Dagstuhl – Leibniz Center for Informatics. It is free to use, not commercial, curated, and of high quality.