local moongoo =require("resty.moongoo")
local cbson =require("cbson")
local mg, err = moongoo.new("mongodb://user:password@hostname/?w=2")
ifnot mg thenerror(err)
endlocal col = mg:db("test"):collection("test")
-- Insert documentlocal ids, err = col:insert({ foo ="bar"})
-- Find documentlocal doc, err = col:find_one({ foo ="bar"})
print(doc.foo)
-- Update documentlocal doc, err = col:update({ foo ="bar"}, { baz ="yada"})
-- Remove documentlocal status, err = col:remove({ baz ="yada"})
-- Close connection or put in OpenResty connection pool
mg:close()
NOTE
You should use cbson datatypes for anything other than strings, floats and bools.
All lua numbers are stored as floats.
Empty arrays are treated and stored as empty documents (you can use cbson.array() to forcibly store empty array in mongo).
nil lua values are ignored and not stored, due to nature of lua, use cbson.null.
nil values from mongo are decoded as cbson.null (use somevar == cbson.null() for comparison)
Moongoo methods
new
<moongoo_obj>mgobj, <string>error = moongoo.new(<string>connection_string)
Creates new Moongoo instance
Refer to Connection String URI Format
Currently supported options are:
w - default is 0
wtimeoutMS - default is 1000
journal - default is false
authMechanism - default depends on mongodb version
socketTimeoutMS - default is nil. It controls both connect and read/write timeout.
It's set to nil by default so you can control connect and read/write separately
with OpenResty lua_socket_connect_timeout, lua_socket_send_timeout and lua_socket_read_timeout.
ssl - default is false
Use ssl for connection. Currently there's no support for CA or client certificates.
Moongoo tries really hard to be smart, and opens connection only when needed,
searches for master node in replicaset, and uses relevant to mongodb version auth mechanism.
Downside is: you can't currently issue queries to slave nodes.
close
mgobj:close()
Closes mongodb connection (LuaSocket) or puts it in connection pool (OpenResty)
Issuing new read/write commands after will reopen connection.
db
<database>dbobj = mgobj:db(<string>name)
Selects database to use.
<collection>colobj = dbobj:collection(<string>name)
Returns new collection object to use.
gridfs
<gridfs>gridfsobj = dbobj:gridfs(<optional string>prefix)
Returns new gridfs object to use.
Default prefix id 'fs'.
cmd
<document>result, <string>error = dbobj:cmd(<string or table>command, <table>params)
Runs database commans.
command is either string with command name, or table { command = value }.
Params are command parameters.
For example, given distinct mongodb command:
<array>result, <string>error = colobj:index_information()
Returns info about current indexes.
ensure_index
<bool>result, <string>error = colobj:ensure_index(<array>indexes)
Creates new index.
indexes should be an array, even if it has 1 value..
Refer to mongo docs for index format.
Note, that in moongoo, index names are optional, moongoo will create them for you based on keys.
drop_index
<bool>result, <string>error = colobj:drop_index(<string>index)
Drops named index from collection.
Refer to mongo docs for index format.
find
<cursor>cursorobj = colobj:find(<table or cbson.oid>query, <table>fields)
Returns new cusor object for query.
find_one
<document>doc, <string>error = colobj:find_one(<table or cbson.oid>query, <table>fields)
Returns first document conforming to query.
find_and_modify
<document>doc, <string>error = colobj:find_and_modify(<table or cbson.oid>query, <table>opts)
Modifies document, according to opts
and returns (by default) old document.
insert
<array>ids, <string or cbson.uint>error_or_number = colobj:insert(<array or table>docs)
Inserts new document(s) and returns their id's and number of inserted documents.
update
<cbson.uint>number, <string>error = colobj:update(<table or cbson.oid>query, <table>update, <table>flags)
Updates document, according to query and flags.
Supported flags are:
multi - update multiple documents (default - false)
upsert - insert document, if not exists (default - false)
Returns number of updated documents
remove
<cbson.uint>number, <string>error = colobj:remove(<table or cbson.oid>queryquery, <bool>single)
Removes document(s) from database.
Returns number of removed documents.
save
<cbson.oid>id = colobj:save(<table>document)
Saves document to collection.
Basically, this performs update with upsert = true, generating id if it not exist.
See here for explanation.
map_reduce
<document>doc, <string>error = colobj:map_reduce(<string>map, <string>reduce, <table>flags) <collection>new_colobj, <string>error = colobj:map_reduce(<string>map, <string>reduce, <table>flags)
Performs map-reduce operation and returns either document with results,
or new collection object (if map-reduce out flag set to collection name).
aggregate
<document>explain, <string>error = colobj:aggregate(<array>pipeline, <table>opts) <collection>new_colobj, <string>error = colobj:aggregate(<array>pipeline, <table>opts) <cursor>cursor, <string>error = colobj:aggregate(<array>pipeline, <table>opts)
Performs aggregation operation according to pipeline commands.
Returns document, if opts.explain set to true.
Returns collection object, if pipeline has $out command as last stage.
Returns new cursor object otherwise.
<cbson.uint>num, <string>err = gridfsobj:remove(<cbson.oid or self-chosen type>id)
Removes file from GridFS.
Returns number of chunks removed.
find_version
<cbson.oid>id, <string>err = gridfsobj:find_version(<string>name, <number>version)
Find version of file, versions from 0 and upwards point to a specific version,
while version from -1 and downwards point to the most recently added version.
E.g. for file with 3 versions: 0, 1, 2 = -3, -2, -1
0 or -3 is oldest, while -1 or 2 - newest.
open
<gridfs.file>gridfsfile, <string>err = gridfsobj:open(<cbson.oid or self-chosen type>id)
Opens GridFS file for reading.
create
<gridfs.file>gridfsfile = gridfsobj:create(<string>filename, <table>opts, <bool>safe)
Creates new GridFS file for writing.
If safe is true (default), all chunks will be inserted only when you call gridfsfile:close().
If safe is false, chunks will be inserted into db with every gridfsfile:write(...),
last chunk will be inserted on close.
You must call gridfsfile:close(), or you'll end up with orphaned chunks.
Safe mode is good for small files, however, as it stores entire file in memory, it's bad for big files.
Non-safe mode uses maximum of (chunkSize*2-1) bytes for any file.
As a side effect, you can :read() or :slurp() file (except for last chunk).
客服电话
APP下载
官方微信
