Hi, If you know all possible authorizations in advance, then you can grant those to a particular user. You may be able to write and use a custom Authorizations security provider, which ensures a user always has every authorization encountered in a visibility string while visibility labels are being parsed for filtering.
A major compaction iterator is a better choice than the scanning API. Iterators applied at the major compaction scope have access to all the underlying data, unfiltered by visibilities, and it would be capable of generating indexes of keys in parallel... but you'd have to rely on some external mechanism for aggregating those indexes from all the compactions across the table. However, if all you want to do is see if entries have changed over time, there's a much simpler way to do that. You can rely on the timestamp/versioning field of Accumulo, to allow multiple versions over time. You may be able to write a Combiner that aggregates the different versions into a single version which tracks how many times it has been changed. Another option would be to use something like Apache Fluo, in your ingest, so that you can incrementally update the counts for how often an entry is modified, during ingest. That's the kind of thing Apache Fluo was designed for. On Mon, Dec 11, 2017 at 8:14 PM Edward Armes <edward.ar...@gmail.com> wrote: > Hi there, > > I was wondering if it was possible in accumulo to retrieve every key in > the table regardless of the viability and classifiers, via the Java API. If > not would it be possible via an iterator? . The idea here would be to build > an index of the keys in accumulo to see when a record is changed over time > in a given accumulo table. > > Thanks >
