A Mongo like shell for NeDB
This shell currently does not hot reload with modifications made by external applications. For example If you have two shells open and edit the database in one shell this will not be reflected in the other. The same goes for having an application running along with the NeDB-Shell, you will have to restart the shell to see the changes made. A file watcher will be added later but currently am working on another project.
Install NodeJs ^6.9.x
Install NeDB
NeDB-Shell has the ability to use all the NeDB functions, for example.
- db.[name].find().exec(cb);
- db.[name].insert().exec();
- db.[name].remove().exec();
And with this package you will have an assortment of new functions on top of the NeDB functions. Check the Table of Contents for all usages.
Install
npm install -g nedb-shellTo begin using NeDB-Shell you will need to navigate to your persistent database directory. If you do not have one yet and would like to create one it would be best to place your directory in AppData on Windows and for mac root. All databases ending with .db in a directory will be loaded into the db object. If there are any files not ending in .db they will not be loaded into the db object for use. Also all created datastores will be created in the current directory with the .db syntax.
Once at your database location simply
> nedb-shellThis will open up a NodeJs shell with this NeDB wrapper module to work in the shell. This way you have all the use of NodeJs shell and NeDB pre-loaded with extra features. This includes auto completion for functions using dot notation.
New additions of 1.1.0, moment and lodash as _ are now pre-loaded.
- Creating/loading a database
- Inserting documents
- Finding documents
- Counting documents
- Updating documents
- Removing documents
- Indexing
- Only on NeDB-Shell
This shell is meant for persistent datastores only. All the operations work on files that are saved to your local machine to persist the database. For reference of the options available to NeDB please refer to the NeDB README. To use NeDB-Shell's create a new datastore in your database directory.
> cd <your database directory>
> nedb-shell
> NeDB Shell: <DISPLAYS CURRENT DATE>
> db.createDatastore('users',{ /* options */ }, //optional cb)
# normal usage
> db.createDatastore('users',{})
# This should simply load a new datastore file into the database.
# Users can now do without the options object.
> db.createDatastore('users')All created datastores are persistent and have the option autoload set to true. The callback is optional and if you decide not to place one there you will automatically receive error and success reporting with an auto set callback.
As for loading the database this is done automatically when you open up NeDB-Shell, if there are any .db files in that directory. When creating a datastore you do not need to append a .db, simply input the name only. When you create a new datastore it is auto loaded into the db object.
#
# Inserting without a callback
> db.users.Insert({ fisrtName: "John", lastName: "Doe"})
# Output
> {
"firstName": "John",
"lastName": "Doe",
"_id": "1234567899876543"
}
#
#
# Inserting a document with a callback
> db.users.Insert({firstName: "John", lastName: "Doe"},function(err,doc){ console.log(doc)});
# Output
> { firstName: 'John', lastName: 'Doe', _id: '1234567899876543' }There is no cursor involved with Insert so you cannot return the document to a variable. However the Find function does use a Cursor and you can use Find to return the value and store it in a variable.
db.[name].Find(query, projection?, cb?), db.[name].FindOne(query,projection?,cb?). Below are some example uses for each find function.
> db.users.Find({ firstName: 'John'})
# As of Version 1.1.0 this no longer logs a cursor.
# Instead it defaults to the cursor function .pretty()
#
# .pretty() - pretty print the documents
# .count() - print count / cannot be chained
# .skip(1).limit(2).sort({_id: -1}) - returns a cursor
# .skip(1).limit(2).sort({_id: -1}).pretty() - prints documents
#
# If you would like to run your own function against the returned cursor doc/docs
# .exec(function(err, docs){
... // docs equals an array
... }
#
# If you would like to return the array of documents simply use the toArray() cursor function
let johnUsers = db.users.Find({ firstName: 'John' }).toArray()
#
#
# There is also FindOne, although this does not return a cursor.
db.users.FindOne({ firstName: 'John' })
> {
"firstName": "John",
"lastName": "Doe",
"_id": "1234567899876543"
}
# or
db.users.FindOne({},function(err,doc){
... // document is an object
... }These find functions are also all available to be uses as they are shown on NeDB if you use the lowercase syntax of find,findOne.
The Count function works just like the Find function but does not return a Cursor and only prints the number found by the query entered.
> db.users.Count()
> 2
> db.users.Insert({ new: "doc" })
> {
"new": "doc",
"_id": "1234567891234567"
}
> db.users.Count()
> 3
> db.users.Count({ new: "doc" })
> 1db.[name].Update(query, updateObject, options, optionalCallBack). Updates only one document unless the multi option is set to true. I added a UpdateMany which has multi set to true automatically if you would rather leave the options empty.
# Update a single document
> db.users.Update({ firstName: "John"}, { firstName: "Phillip" })
> { changed: 1,
affectedDocument: {},
affectedDocuments: [],
upsert: false }
#
# The returned object string is not assignable to a variale but it does show the number
# of documents changed and also if any documents where made because of upsert being true.
# To have either the affectedDocumentor affectedDocuments filled with the _id/s of changed
# documents you have to set `returnUpdatedDocs: true` in the `options`.
#
# Here I will update many and see the changed object _ids
db.users.Update({ firstName: "John"}, { firstName: "Phillip"}, {multi: true,
... returnUpdatedDocs: true})
> { changed: 3,
affectedDocument: {},
affectedDocuments: [ '1234567899876543', '9876543211234567', '1234567891234567' ],
upsert: false }
#
# To print out the full documents you could write your own callback function
> db.users.Update({ firstName: "John"},{firstName: "Phillip"},{multi: true},function(err, docs){
... // docs will be in the form of an array
}db.users.Remove(query, options, callback// optional) will remove one document but if you set multi: true in the options this will remove all documents matching the query.
> db.users.Remove({ firstName: "John"},{ multi: true})
> { removed: 3 }
# prints an object with the number removed.
#
# with custom callback
> db.users.Remove({ firstName: 'John'},{ multi: true}, function(err,numberRemoved) {
... // numberRemoved is of type number
}
>db.users.EnsureIndex(query/optionsObject) is used to create an index on a value, The _id value is already indexed. db.users.RemoveIndex(fieldName, cb)
> db.users.EnsureIndex({ fieldName: 'firstName', unique: true })
> Success
# If there is an error or conflict during indexing an error will show up in red.
> db.users.RemoveIndex('firstName')
> Successdb.[name].FindOneAndUpdate(query,update,updateOptions?,cb?). db.[name].FindOneAndRemove(query,cb?). Both are not in the NeDB library but can be implemented. These were added for users who would like to circumvent writing more code than needed. This also adds another layer of error checking. If the FindOne comes back with an error the Update or Remove will not execute.
# To find one document and update that single document in one query
# query update
> db.users.FindOneAndUpdate({ firstName: 'John' },{ firstName: 'Max' });
> { changed: 1,
affectedDocument: {},
affectedDocuments: [],
upsert: false }
#
# As you can see you still receive the reply document. If you wish you can set to retrieve the id as well.
#
> db.users.FindOneAndUpdate({ firstName: 'John'},{ firstName: 'Max' },{ returnUpdatedDocs: true });
> { changed: 1,
affectedDocument: '1234567899876543',
affectedDocuments: [],
upsert: false }
#
# However You cannot set multi to true. FindOneAndRemove is simply available to view the removed document before hand by using the attribute show: true in the second argument. To simply remove just one document use removing documents. Remove's default multi: false will only remove one document. With FindOneAndRemove you might have an error finding the document and be warned about it.
db.users.FindOneAndRemove({ firstName: 'John' })
> { removed: 1 }
#
# with returned document
#
> db.users.FindOneAndRemove({},{ show: true })
> {
"firstName": "John",
"lastName": "Doe",
"_id": "1234568797894563"
}db.[name].InsertMany(array,cb?). A callback is available and would function on each document in the array. If left blank there is an automatic callback that will log the error and the document it failed on. Great for pasting in an array of objects to insert.
> let bushel = [{ name: 'Chris'},{ name: 'Tom' },{ name: 'Mercedes'}];
> db.users.InsertMany(bushel);
#
# If you would like to write your own callback to log every success
#
> db.users.InsertMany(bushel,function(err,doc){if(!err){
... console.log("success")});
> success
success
successdb.[name].UpdateMany(query,update,updateOptions?,cb?). For when You would rather not have to set multi: true in the options all the time.
# for an override
> db.UpdateMany({ name: 'John'},{ name: 'Sierra' },{ returnUpdateDocs: true})
> { changed: 2,
affectedDocument: {},
affectedDocuments: [ '132424523452345', '234524523423452' ],
upsert: false }
#
# Use update options like $set to not override the entire document.
#
> db.UpdateMany({ name: 'John' }, { name: 'Sierra' })
> { changed: 2,
affectedDocument: {},
affectedDocuments: [],
upsert: false }
#
# Again the _ids are only returned when returnUpdatedDocs is set to true in the options.db.[name].RemoveMany(query,cb?). For when you would rather not have to set the multi: true options query.
> db.users.RemoveMany({name: 'John'})
> { removed: 2 }
# Auto sets multi to true. and outputs the same removed object string.
# Optional callback. Added drop functionality to remove databases from the current directory.
> db.users.Drop();
# removes users.db from current directory. Also removes from db object.
#
> db.Drop();
> let dropList = ['users','locations','cities'];
> db.Drop(dropList);
# Without any array List db.Drop will drop all datastores in the current
# directory.This feature should give you a great place to test code out on the nodejs shell running in ES6. I have added the ability for anyone to load in their own .js files into the context of this shell. These of course have to be either written in or compiled to ES6 or lower javascript modules. Lets say I have my persistent database at ~/data and I have my .js module functions in another directory, lets say ~/scripts. The path /Users/<yourUserName> has to be used in place of ~. So to accesss your scripts from any directory past ~, you will need to pass in that path.
# to load in one singular function into the context
> load.Function('/Users/<yourUserName>/scripts/add_two.js')
# now you can use add_two as a function in the context of the shell.
> add_two(2,2)
> 4
# If you would like to load an entire directory of functions you simply
# point to the directory you would like to load.
> load.Directory('/Users/<yourUserName>/scripts')
#
# Now all the .js files in that directory are loaded into the context.
# To see all the functions you have loaded in simple use show
> load.Show()
add_two
add_five
#Now that you know how to load in the functions you have created, how exactly do these need to look to be loaded in correctly. Simply use nodejs modules in each file or even create a module that loads many modules out of a nested directory.
// Example add_two.js
"use strict";
module.exports = (a,b) => a+b;
// that is is. The name of the function in the context will be the name
// of the file the module is in. So add_two's function will be used as
// add_two(2,2);Last but not least extra functions for help and other interactions not dealing with the datastores.
> db.Help() # list all functions available on the db object.
> load.Help() # list all the functions available on the load object.
> db.[name].Help() # List all functions on the datastore object.
#
# Version 1.1.0 additions clear().
# If you would like to clear the screen
> clear()