•

CRUD/Mongo/SQL terminology

•

Create = Insert = INSERT

•

Read = Find = SELECT

•

Update = Update = UPDATE

•

Delete = Remove = DELETE

•

MongoDB's CRUD operation exist as methods/functions in programming language APIs

•

MongoDB does not have a separate language that needs to be embedded as strings inside code that operates on MongoDB

•

as a developer, you manipulate the database using methods on objects, similar to a object mapping with RDBMS

•

this makes database I/O with MongoDB relatively pleasant for programmers

•

the Mongo shell

•

the Mongo shell is an interactive JavaScript interpreter

•

pay attention to the version number when you log in

•

you can type in javascript at the command line

•

up arrow retrieves previous commands

•

ctrl-a to front of line

•

built-in helpers

•

help

•

help keys

•

command completion

•

basic javascript variable assignment

•

BSON

•

fundamental record type is a nested dictionary of key/value associations

•

javascript objects, because they are in dictionary/JSON form, can be saved directly into the database

•

JSON was inspired by the syntax of JavaScript

•

MondoDB does not use this stringy syntax

•

http://bsonspec.org ; BSON spec ; BSON is binary JSON, has data types that go beyond JSON, e.g. datetime, 32 and 64 bit integer type, etc.

•

in the shell we use JavaScript and we can force the types of numbers

•

NumberInt(1)

•

NumberLong(1)

•

but be careful with languages such as JavaScript or Perl which can't represent all of the types that BSON can represent

•

new Date() or ISODate(...)

•

so you can faithly handle all of the data that comes out of the database by casting it like this:

•

inserting docs

•

the shell has a variable named db which is a handle to the current database

•

collections are properties of the database, e.g. db.people.insert(doc)

•

when you insert a document, it gets a unique field "_id"

•

different from other databases, the primary key is immutable

•

you could similate changing the id by removing a document, and inserting it again so it gets another id

•

_id is the primary key

•

_id is always ObjectId made up of date, machine, process id on computer, and a counter, so these ideas are globally unique

•

if you don't insert an _id, then one will be generated for you

•

finding documents

•

db.fruit.find() will find all

•

•

the query is presented to the server in the form of a structured document

•

the second argument allows you to specify which fields you want to get back, true or false

•

_id is included by default

•

find

•

fill database so we can do some queries:

•

db.people.find() to see them all

•

type it iterate through the results

•

shell is keeping these results open

•

cursor is on the server, will clear it in 10 minutes

•

dp.people.find().pretty() to see results formatted

•

when multiple fields are given in the first parameters of find(), then it assumes AND

•

these are called "query by example" queries

•

query operators

•

db.scores.find({ score : { $gt : 95, $lte : 95}) (greater-than, less-than-or-equal)

•

&gt, &lt etc. can also be applied to strings

•

currently sorts asciibetically

•

future versions should have better support

•

all comparison operations are strongly typed

•

if you filter by less than "A", you will not get numbers

•

generally, you should not mix types in one named field, e.g. store 42 in a name

•

here the third item will not work since comparisons are case sensitive

•

querying on whether field exists

•

you can check for a type, use the BSON specifications, so string = 2:

•

•

you can also use regexs:

•

all names that end with "e"