In the case when index is of type {_}, index is returned as is
instead of putting it into a tuple {Ix, ordered}.
An example
Given the following index plugin implementation:
-module(key6).
-record(cgh, {key, val}).
-export([k6/3
, add_index/0
, create_table/0
, tst_bad_type/0
]).
k6(_, _, #cgh{key = {Name, K8, _Cid, _Vn}}) ->
[list_to_binary([atom_to_binary(Name),
list_to_binary(lists:sublist(K8, 6))])].
add_index() ->
mnesia_schema:add_index_plugin({k6}, key6, k6).
create_table() ->
mnesia:create_table(cgh,
[{attributes, record_info(fields, cgh)},
{type, ordered_set},
{index, [{k6}]},
{rocksdb_copies, [node()]},
{local_content, true}]).
tst_bad_type() ->
E1 = #cgh{key = {n1, "u1vgrjkh",{ecgi,238,6,213020,50},3}, val = ok},
mnesia:dirty_write(E1),
Index1 = list_to_binary([atom_to_binary(n1),
list_to_binary(lists:sublist("u1vgrjkh", 6))]),
mrdb:index_read(cgh, Index1, {k6}).
-- testing --
Eshell V14.2.2 (press Ctrl+G to abort, type help(). for help)
(rock@bang)2> mnesia:create_schema([node()]).
ok
(rock@bang)3> mnesia:start().
ok
(rock@bang)4> mnesia_rocksdb:register().
{ok,rocksdb_copies}
(rock@bang)5> l(key6).
{module,key6}
(rock@bang)6> key6:add_index().
{atomic,ok}
(rock@bang)7> key6:create_table().
{atomic,ok}
(rock@bang)8> key6:tst_bad_type().
** exception exit: {aborted,{bad_type,{cgh,index,{k6}}}}
in function mnesia:abort/1 (mnesia.erl, line 362)
in call from mrdb:ensure_ref/1 (~/projects/erlang/github/mnesia_rocksdb/src/mrdb.erl, line 609)
in call from mrdb:index_read_/3 (~/projects/erlang/github/mnesia_rocksdb/src/mrdb.erl, line 945)
-- after patch --
(rock@bang)9> c(mrdb).
Recompiling ~/projects/erlang/github/mnesia_rocksdb/src/mrdb.erl
{ok,mrdb}
(rock@bang)10> key6:tst_bad_type().
[{cgh,{n1,"u1vgrjkh",{ecgi,238,6,213020,50},3},ok}]
(rock@bang)11>
Expose low-level helpers, fix dialyzer warnings
WIP column families and mrdb API
Basic functionality in place
started adding documentation
remove doc/ from .gitignore
add doc/* files
recognize pre-existing tabs at startup
wip: most of the functionality in place (not yet merge ops)
wip: adding transaction support
wip: add transaction test case (currently dumps core)
First draft, mnesia plugin user guide
Fix note formatting
WIP working on indexing
Index iterators, dialyzer, xref fixes
open db with optimistic transactions
Use rocksdb-1.7.0
Use seanhinde rocksdb patch, enable rollback
Call the right transaction_get() function
WIP add 'snap_tx' activity type
tx restart using mrdb_mutex
Fix test suite sync bugs
WIP instrumented for debugging
WIP working on migration test case
Add migration test suite
Migration works, subscribe to schema changes
WIP fix batch handling
Manage separate batches per db_ref
Add mrdb:fold/3
Add some docs, erlang_ls config
Use seanhinde's rocksdb vsn
During initialization of the backend the caller may provide a new
configuration to define a target ETS table for storing errors which the
backend encountered at runtime. These errors are written during handling
of write errors.
- put(), write() and delete() wrapped, checking for non-ok returns
- Option {on_write_error, verbose | warning | error | fatal}
added to the rocksdb_opts user property. The corresponding function
mnesia_lib:Op(Fmt, Args) will be called if an error return is spotted.
first select clause in a multi-clause select pattern. This could give
erroneous results if the first pattern was more specific (on the key)
than following patterns. This change calculates a prefix for all clauses
and keeps the shortest common prefix.
