Skip to content

Record based Erlang driver for MongoDB with gridfs support

License

Notifications You must be signed in to change notification settings

b00ris/erlmongo

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

98 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Info

Erlmongo is a pretty complete Erlang driver for mongodb.

It supports records and proplists as datatypes. Strings can be lists or binaries, but strings received from mongodb (as a result of find) will be binaries.

Because of the way records work in Erlang, you need to call mongoapi:recinfo/2 before using any record, or define each record in erlmongo.hrl.

When you’re using a selector (picking which fields in the document you wish to get from mongodb), they have to be in the same sequence as they were defined in the record. For instance:

% -record(mydoc {name, i}).
% This will work
Mong:findOne(#mydoc{i = 10}, [#mydoc.name, #mydoc.i]).
% This will NOT work
Mong:findOne(#mydoc{i = 10}, [#mydoc.i, #mydoc.name]).

WARNING: If you use modifier updates with {set,Value} ($set), do not use records if that value is a string that changes size (it should work for integers for instance). Mongo changes the document element sequence and this messes up record serialization.

Connections are always single to the master:

  • master/slave - read and write from master

  • master/master - pick a master at random and use it for everything

  • replica pairs/sets - find out which is master and connect to it

If your usage pattern is read heavy and want to use slaves for reading, connect to them with singleServer call. Don’t write to slave connections or bad things will happen.

Always use an atom for naming connections. The connection process will register itself on that name. Runtime connection API: mongodb:singleServer/1,2 mongodb:replicaPairs/3 mongodb:replicaSets/2 mongodb:masterSlave/3 mongodb:connect/1 mongodb:sharded/2

Connection info is saved in an application variable (erlmongo,connections) and updated with every call to: singleServer, masterSlave, replicaPairs, masterMaster, replicaSets, sharded and deleteConnection. If mongodb process crashes for some reason, connections will be restarted. Also if the application itself is stopped and started. You can add a connections variable to erlmongo.app file:

% def is name of connection, types can be: masterSlave,masterMaster or replicaPairs. This is how singleServer info is saved:
{env, [{def,{masterSlave,{"localhost",27017},{"localhost",27017}}}]},

Replica sets

% List of servers does not have to be the entire list of the replica set.
% Erlmongo will read the primary server from them and connect to it (even if not in the list).
mongodb:replicaSets(repl,["127.0.0.1:30000","127.0.0.1:30001"]).
mongodb:connect(repl).

Examples

make
erl
rr("erlmongo.hrl").
application:start(erlmongo).
% Set mongodb server info. singleServer(PoolName) is the same as singleServer(PoolName,"localhost:27017")
mongodb:singleServer(def).
mongodb:connect(def).
% Create an interface for test database (it has to be a binary)
Mong = mongoapi:new(def,<<"test">>).

% Save a new document
Mong:save(#mydoc{name = "MyDocument", i = 10}).
% Return the document, but only the "i" field (+ _id which always gets returned)
Mong:findOne(#mydoc{i = 10}, [#mydoc.name]).

% With proplists
Mong:save("mydoc", [{"name", "MyDocument"}, {"i", 10}]).
Mong:findOne("mydoc", [{"i", 10}], [{"name", 1}]).

% Switch to mochijson's style proplists (string = binary, array = list, no list-coded strings)
Mong:set_encode_style(mochijson),
Mong:save("mydoc", [{<<"name">>, <<"2-dim array">>}, {<<"data">>, [[1.1, 2.2], [3.3, 4.4]]}]).

% Set Index. First parameter is so that the driver knows what collection
%  we mean. If you have an already constructed record laying around use that.
%  No need to construct a new record just so the driver can read the name.
% Second parameter the index we wish to create. 1 = ascending, -1 = descending.
Mong:ensureIndex(#mydoc{}, [{#mydoc.i, 1}, {#mydoc.name, -1}])

% Find examples:

% Or
M:find("mydoc",[{'or',[{"a",1},{"i",11}]}],undefined,0,100).

% Parameters: Search criteria, field selector, docs to skip, docs to return
Mong:find(#mydoc{i = 4}, [#mydoc.name], 0, 0).
% Same thing but with #search record that provides default parameters
Mong:find(#search{criteria = #mydoc{i = 4}, field_selector = [#mydoc.name]}).

% Find with options
Mong:findOpt(#mydoc{i = 4}, undefined, [explain], 0, 0).
% Same thing as above
Mong:findOpt(#search{criteria = #mydoc{i = 4}}, [explain]).
% Also the same, with proplists
Mong:findOpt("mydoc", #search{criteria = [{"i",  4}]}, [explain]).

% Embedded records
Mong:save(#mydoc{name = "zembedom", i = 10, address = #address{city = "ny", street = "some", country = "us"}}).
Mong:find(#mydoc{address = #address{city = "la"}}, undefined, 0, 0).

% Advanced queries (supported: gt, lt, gte, lte, ne, in, nin, all, size, exists):
% Documents with even i
Mong:find(#mydoc{i = {mod, 2, 0}}, undefined, 0,0).
% Documents with i larger than 2:
Mong:find(#mydoc{i = {gt, 2}}, undefined, 0,0).
% Documents with i between 2 and 5:
Mong:find(#mydoc{i = {in, {gt, 2}, {lt, 5}}}, undefined, 0,0).
% in example:
Mong:find(#mydoc{tags = {in, [2,3,4]}}, undefined, 0,0).
% exists example:
Mong:find(#mydoc{tags = {exists, false}}, undefined, 0,0).

% findandmodify command
Mong:runCmd([{"findandmodify", "collectionname"},{"query", [{"fn","onmeta.flv"},{"ch","somechan"}]},{"remove",1}]).

% GridFS
% Always run this on collection before writing the first file
Mong:gfsIndexes().
{ok, Bin} = file:read_file("SomeFile").
% To open file for writing, use gfsNew
PID = Mong:gfsNew("myfile").
% You can set parameters: mime, meta (embedded document), aliases (array of names), chunk size (default 256k)
%                         flushLimit (at which buffer size data gets flushed to mongodb, def. 1MB)
% PID = Mong:gfsNew("myfile", [{chunkSize, 100}]).
% You can also set collection name (default is fd)
% PID = Mong:gfsNew("myfilecol", "myfile", []).
Mong:gfsWrite(PID,Bin).
Mong:gfsClose(PID).
% Reading
PID = Mong:gfsOpen(#gfs_file{filename = "myfile"}).
Res = Mong:gfsRead(PID,100000).
Mong:gfsClose(PID).

Supported Data types and modifiers

Look at mongodb:encode_element/1

Switching to mochijson’s style proplists

Mochijson library has a different string/list encoding convention: string = binary array = list You can choose a preferred way to encode with mongoapi:set_encode_style/1. Selected style is set to a selected server and table and not for mongoapi module instance. An example:

...
12> Mong:set_encode_style(default).
ok
13> Mong:save("foobar", [{<<"data">>, [[1.1, 2.2], [3.3, 4.4]]}]).
** exception error: bad argument
     in function  unicode:characters_to_binary/1
        called as unicode:characters_to_binary([[1.1,2.2],[3.3,4.4]])
     in call from mongodb:encode_cstring/1
     in call from mongodb:encode_element/1
     in call from mongodb:'-encode/2-fun-0-'/3
     in call from lists:foldl/3
     in call from mongodb:encode/2
     in call from mongoapi:save/3
14> Mong:set_encode_style(mochijson).
ok
15> Mong:save("foobar", [{<<"data">>, [[1.1, 2.2], [3.3, 4.4]]}]).
{oid,<<"000af08b902dee723e000006">>}

Supported operation list

Collections

  • remove

  • save

  • insert

  • update

  • batchInsert

  • ensureIndex

  • deleteIndex

  • deleteIndexes

  • count

  • dropCollection

  • createCollection

  • group

Search

  • find

  • findopt

  • cursor - getMore - closeCursor

  • findOne

DB

  • eval

  • stats

  • runCmd

  • repairDatabase

  • cloneDatabase

  • dropDatabase

  • addUser

  • setProfilingLevel

  • getProfilingLevel

GridFS

  • gdsIndexes

  • gfsNew

  • gfsWrite

  • gfsOpen

  • gfsRead

  • gfsDelete

  • gfsFlush

  • gfsClose

Building

erl -make

or with autotools ./configure make make install

Author

Sergej Jurečko

About

Record based Erlang driver for MongoDB with gridfs support

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published