Using MongoDB and Python - Demo

2/23/2016 2016Feb23_Pyugre_MongoDB_DEMO
http://localhost:8888/notebooks/2016Feb23_Pyugre_MongoDB_DEMO.ipynb# 1/32
The Mongo Shell
PyMongo
Accessing the mongo shell
TOP
This section shows usage of the MongoDB shell, but due to the noninteractive nature of the Jupyter
Notebook cells, we show each interaction as a commandline invocation of the "mongo" shell client
with the command to execute as an argument.
We start by defining an IPython alias to facilitate running mongo commands.
In [1]:
NOTE: To use mongo alias, must be in quotes and only one statement (last one) is executed
NOTE: We must always give the name of a database (even one that doesn't exist) to be able to run a
query
In [2]:
Let's get a list of collections contained within the 'Money_UK' database:
In [3]:
MongoDB shell version: 3.2.3
connecting to: localhost:27017/ANY
3.2.3
connecting to: localhost:27017/Money_UK
[ "CARD", "SAVING" ]
%alias mongo c:/Progs/MongoDB/bin/mongo.exe ‐‐eval %s localhost:27017/%s
mongo "db.version()" ANY
mongo "db.getCollectionNames()" Money_UK

In [4]:
Let's see what commands are available in the MongoDB shell
{
"db" : "Money_UK",
"collections" : 2,
"objects" : 2974,
"avgObjSize" : 161.98520511096166,
"dataSize" : 481744,
"storageSize" : 192512,
"numExtents" : 0,
"indexes" : 2,
"indexSize" : 53248,
"ok" : 1
}
mongo "db.stats()" Money_UK

In [5]:
mongo "db.help()" Money_UK

DB methods:
db.adminCommand(nameOrDocument) ‐ switches to 'admin' db, and r
uns command [ just calls db.runCommand(...) ]
db.auth(username, password)
db.cloneDatabase(fromhost)
db.commandHelp(name) returns the help for the command
db.copyDatabase(fromdb, todb, fromhost)
db.createCollection(name, { size : ..., capped : ..., max : ...
} )
db.createUser(userDocument)
db.currentOp() displays currently executing operations in the d
b
db.dropDatabase()
db.eval() ‐ deprecated
db.fsyncLock() flush data to disk and lock server for backups
db.fsyncUnlock() unlocks server following a db.fsyncLock()
db.getCollection(cname) same as db['cname'] or db.cname
db.getCollectionInfos([filter]) ‐ returns a list that contains
the names and options of the db's collections
db.getCollectionNames()
db.getLastError() ‐ just returns the err msg string
db.getLastErrorObj() ‐ return full status object
db.getLogComponents()
db.getMongo() get the server connection object
db.getMongo().setSlaveOk() allow queries on a replication slave
server
db.getName()
db.getPrevError()
db.getProfilingLevel() ‐ deprecated
db.getProfilingStatus() ‐ returns if profiling is on and slow t
hreshold
db.getReplicationInfo()
db.getSiblingDB(name) get the db at the same server as this one
db.getWriteConcern() ‐ returns the write concern used for any o
perations on this db, inherited from server object if set
db.hostInfo() get details about the server's host
db.isMaster() check replica primary status
db.killOp(opid) kills the current operation in the db
db.listCommands() lists all the db commands
db.loadServerScripts() loads all the scripts in db.system.js
db.logout()
db.printCollectionStats()
db.printReplicationInfo()
db.printShardingStatus()
db.printSlaveReplicationInfo()
db.dropUser(username)
db.repairDatabase()
db.resetError()
db.runCommand(cmdObj) run a database command. if cmdObj is a s
tring, turns it into { cmdObj : 1 }
db.serverStatus()
db.setLogLevel(level,<component>)
db.setProfilingLevel(level,<slowms>) 0=off 1=slow 2=all
db.setWriteConcern( <write concern doc> ) ‐ sets the write conc
ern for writes to the db

Let's see how many documents exist in the CARD collection:
In [6]:
Let's see how many documents exist in the SAVING collection:
In [7]:
Let's show the 'first' document in the CARD collection:
In [8]:
Let's search for any CARD documents with that same timestamp:
db.unsetWriteConcern( <write concern doc> ) ‐ unsets the write
concern for writes to the db
db.setVerboseShell(flag) display extra information in shell out
put
db.shutdownServer()
db.stats()
db.version() current version of the server
1493
1481
{
"_id" : ObjectId("56cb51c37469f7afbc7f375b"),
"date" : ISODate("1988‐10‐21T00:00:00Z"),
"value" : 18.53,
"info" : "Interest",
"account" : "./UK_GBP_Access‐CC",
"tags" : [
"interest",
"interest"
],
"comment" : ""
}
mongo "cc=db.CARD;print(cc.find().count())" Money_UK
mongo "saving=db.SAVING;print(saving.find().count())" Money_UK
mongo "cc=db.CARD;cc.findOne()" Money_UK

In [9]:
Let's perform the same search but provide a projection to show only the value field
NOTE: it will show the '_id' index field unless we explicitly disable it.
In [10]:
Now let's perform the same search but provide a projection to show only the value field, without the
'_id' index field
In [11]:
We can also perform the same search and sort the result in ASCENDING(1) or DESCENDING(1)
order:
In [12]:
{ "_id" : ObjectId("56cb51c37469f7afbc7f375c"), "date" : ISODate("1988‐
09‐27T00:00:00Z"), "value" : 20.4, "info" : "Filton Tech. Coll.", "acco
unt" : "./UK_GBP_Access‐CC", "tags" : [ "misc", "unknown" ], "comment"
: "" }
{ "_id" : ObjectId("56cb51c37469f7afbc7f375e"), "date" : ISODate("1988‐
09‐27T00:00:00Z"), "value" : 29.97, "info" : "Burtons, Glasgow", "accou
nt" : "./UK_GBP_Access‐CC", "tags" : [ "bills", "clothes" ], "comment"
: "" }
{ "_id" : ObjectId("56cb51c37469f7afbc7f375c"), "value" : 20.4 }
{ "_id" : ObjectId("56cb51c37469f7afbc7f375e"), "value" : 29.97 }
{ "value" : 20.4 }
{ "value" : 29.97 }
{ "value" : 20.4 }
{ "value" : 29.97 }
mongo "cc=db.CARD;cc.find({'date':ISODate('1988‐09‐27T00:00:00Z')})" Money_UK
mongo "cc=db.CARD;cc.find({'date':ISODate('1988‐09‐27T00:00:00Z')},{value:1})" Money_UK
mongo "cc=db.CARD;cc.find({'date':ISODate('1988‐09‐27T00:00:00Z')},{_id:0,value:1})"
mongo "cc=db.CARD;cc.find({'date':ISODate('1988‐09‐27T00:00:00Z')},{_id:0,value:1}).sort({v

In [13]:
In the following example we show many entries, but the output is automatically limited to 20
documents:
In [14]:
Here, we limit the output to only 5 documents:
In [15]:
{ "value" : 29.97 }
{ "value" : 20.4 }
{ "value" : ‐1695.61 }
{ "value" : ‐1200 }
{ "value" : ‐1000 }
{ "value" : ‐834.51 }
{ "value" : ‐715.46 }
{ "value" : ‐703.16 }
{ "value" : ‐670 }
{ "value" : ‐652.37 }
{ "value" : ‐600 }
{ "value" : ‐590 }
{ "value" : ‐535 }
{ "value" : ‐500 }
{ "value" : ‐490.21 }
{ "value" : ‐316.81 }
{ "value" : ‐238.42 }
{ "value" : ‐210 }
{ "value" : ‐200 }
{ "value" : ‐184.25 }
{ "value" : ‐150 }
{ "value" : ‐150 }
Type "it" for more
{ "value" : ‐1695.61 }
{ "value" : ‐1200 }
{ "value" : ‐1000 }
{ "value" : ‐834.51 }
{ "value" : ‐715.46 }
mongo "cc=db.CARD;cc.find({'date':ISODate('1988‐09‐27T00:00:00Z')},{_id:0,value:1}).sort({v
mongo "cc=db.CARD;cc.find({},{_id:0,value:1}).sort({value:1})" Money_UK
mongo "cc=db.CARD;cc.find({},{_id:0,value:1}).sort({value:1}).limit(5)" Money_UK

We can also specify to skip the first 5 documents (it will show the following 20 documents):
In [16]:
SQL Equivalent
The equivalent query in SQL, for the query below, would be:
SELECT value            # Projection
FROM cc                 # Table (Collection)
WHERE value > ‐100.00   # select criteria
LIMIT 5                 # cursor modifier
Below we show the 1st five values with value less than 100.00
In [17]:
or the info and value fields associated with all 'media_music' tags:
{ "value" : ‐703.16 }
{ "value" : ‐670 }
{ "value" : ‐652.37 }
{ "value" : ‐600 }
{ "value" : ‐590 }
{ "value" : ‐535 }
{ "value" : ‐500 }
{ "value" : ‐490.21 }
{ "value" : ‐316.81 }
{ "value" : ‐238.42 }
{ "value" : ‐210 }
{ "value" : ‐200 }
{ "value" : ‐184.25 }
{ "value" : ‐150 }
{ "value" : ‐150 }
{ "value" : ‐124.94 }
{ "value" : ‐108.18 }
{ "value" : ‐107.84 }
{ "value" : ‐104.46 }
{ "value" : ‐104 }
Type "it" for more
{ "value" : ‐85.58 }
{ "value" : ‐79.8 }
{ "value" : ‐76.15 }
{ "value" : ‐63.16 }
{ "value" : ‐50.1 }
mongo "cc=db.CARD;cc.find({},{_id:0,value:1}).sort({value:1}).skip(5)" Money_UK
mongo "cc=db.CARD;cc.find({value:{$gt:‐100.00}},{_id:0,value:1}).sort({value:1}).limit(5)"

In [18]:
Using the Pymongo driver
TOP
Now let's look at using the Pymongo driver to access our 'Money_UK' database.
Let's first connect to MongoDB, creating a 'client' object and then list the available databases:
In [19]:
Now let's explicitly connect to the 'Money_UK' DB and then list the collections contained within it:
{ "value" : 27.41, "info" : "Sing Chong PTE Ltd." }
{ "value" : 23.25, "info" : "CD Cases Maplin" }
{ "value" : 18, "info" : "Sunrise Records, Toronto" }
{ "value" : 21.29, "info" : "Musique D"Auteuil, Quebec" }
{ "value" : 29.5, "info" : "Tower Records, NY" }
{ "value" : 20.84, "info" : "The Wiz/New York" }
{ "value" : 16.87, "info" : "Records on Wheels, Ottawa" }
{ "value" : 65.92, "info" : "HMV,Manchester" }
{ "value" : 44.56, "info" : "French CDs,EuroM" }
{ "value" : 12.57, "info" : "CD,La FNAC{Paris, 121 F}" }
{ "value" : 15.37, "info" : "CD,film" }
{ "value" : 18.2, "info" : "CDs{Slade/J Winter, }" }
{ "value" : 36.67, "info" : "Carrefour CDs" }
{ "value" : 38.39, "info" : "La FNAC CDs" }
{ "value" : 33.29, "info" : "HMV CDs" }
{ "value" : 39, "info" : "Hi‐Fi News/RR Subscription,Nov92‐Jan94" }
{ "value" : 31.96, "info" : "HMV Oxford St." }
{ "value" : 25.68, "info" : "Carrefour CDs" }
{ "value" : 21.93, "info" : "Hi‐Fi News/RR CD Service" }
{ "value" : 8.61, "info" : "Hi‐Fi News/RR CD Service" }
Type "it" for more
Available databases: <['Money_UK', 'aggregation_example', 'local', 'tes
t']
mongo "cc=db.CARD;cc.find({'tags':['media','music']},{_id:0,info:1,value:1})" Money_UK
# First install pymongo client:
# conda install pymongo
from pymongo import MongoClient,ASCENDING,DESCENDING
client = MongoClient("mongodb://localhost:27017")
dbnames = [dbname for dbname in client.database_names() ]
print("Available databases: <" + str(dbnames))

In [20]:
Below we perform a find() operation on the 'CARD' and 'SAVING' collections.
The find() method returns a cursor on which we can iterate.
We use the cursor to show just the first N(4) documents in each collection
Opened db<Money_UK>
Available collections in db<Money_UK>: ['CARD', 'SAVING']
db = client.Money_UK
dbname = db.name
print("Opened db<" + dbname + ">")
collnames = [collname for collname in db.collection_names() ]
print("Available collections in db<" + dbname + ">: " + str(collnames))

In [21]:
Number of documents in CARD collection: 1493
1st 4 documents:
{'account': './UK_GBP_Access‐CC', 'comment': '', '_id': ObjectId('56cb5
1c37469f7afbc7f375b'), 'date': datetime.datetime(1988, 10, 21, 0, 0),
'tags': ['interest', 'interest'], 'info': 'Interest', 'value': 18.53}
1c37469f7afbc7f375c'), 'date': datetime.datetime(1988, 9, 27, 0, 0), 't
ags': ['misc', 'unknown'], 'info': 'Filton Tech. Coll.', 'value': 20.4}
1c37469f7afbc7f375d'), 'date': datetime.datetime(1988, 9, 1, 0, 0), 'ta
gs': ['bills', 'house'], 'info': 'Flowers, W_O_T', 'value': 8.0}
1c37469f7afbc7f375e'), 'date': datetime.datetime(1988, 9, 27, 0, 0), 't
ags': ['bills', 'clothes'], 'info': 'Burtons, Glasgow', 'value': 29.97}
Number of documents in SAVING collection: 1481
1st 4 documents:
{'account': './UK_GBP_Barclays_CA', 'comment': '', '_id': ObjectId('56c
b51c47469f7afbc7f3ae3'), 'date': datetime.datetime(1988, 9, 27, 0, 0),
'tags': ['misc', 'pay'], 'info': 'ACCESS', 'value': ‐1200.0}
'tags': ['bills', 'house'], 'info': 'Mortgage', 'value': ‐317.9}
'tags': ['bills', 'house'], 'info': 'Rates', 'value': ‐58.4}
'tags': ['bills', 'house'], 'info': 'Endowment', 'value': ‐31.8}
# Query for All Documents in a Collection
# To return all documents in a collection, call the find() method without a criteria docume
# For example, the following operation queries for all documents in the restaurants collect
CARD = db.CARD
SAVING = db.SAVING
cursor = CARD.find()
print()
print("Number of documents in CARD collection: " + str( cursor.count() ))
N=4
print("1st {} documents:".format(N))
for document in cursor.limit(N):
print(document)
cursor = SAVING.find()
print()
print("Number of documents in SAVING collection: " + str( cursor.count() ))
N=4
print("1st {} documents:".format(N))
print(document)

We can also apply sorting the find() call, which also returns the cursor to use.
In the following example we see the effect of using ASCENDING or DESCENDING sort (in the
descending case we see entries from year 2013).

In [23]:
Below we specify an empty match condition '{}' on the find() call, but we also provide a '{value: 1, _id:
0}' projection so that this cursor only returns the value field for each document.
1st 4 documents: ORDERED by ASCENDING date
b51c47469f7afbc7f3af2'), 'date': datetime.datetime(1988, 9, 1, 0, 0),
'tags': ['fun', 'hols'], 'info': 'Borth YHA', 'value': ‐10.0}
'tags': ['misc', 'pay'], 'info': 'ACCESS', 'value': ‐1200.0}
'tags': ['salary', 'salary'], 'info': 'Wage', 'value': 883.65}
'tags': ['bills', 'house'], 'info': 'Rates', 'value': ‐58.4}
1st 4 documents: ORDERED by DESCENDING date
{'account': './UK_GBP_FirstD_ESA', 'comment': '', '_id': ObjectId('56cb
51c57469f7afbc7f416d'), 'date': datetime.datetime(2013, 1, 31, 0, 0),
'tags': ['interest', 'int'], 'info': 'Interest', 'value': 0.1}
51c57469f7afbc7f416a'), 'date': datetime.datetime(2013, 1, 21, 0, 0),
'tags': ['transfer', 'transfer'], 'info': 'Transfer to France, Axa{for
Prestation Compensatoire}', 'value': ‐6520.0}
51c57469f7afbc7f416c'), 'date': datetime.datetime(2013, 1, 21, 0, 0),
'tags': ['bills', 'bank'], 'info': 'Cost Of Transfer{for Prestation Com
pensatoire}', 'value': ‐9.0}
{'account': './UK_GBP_FirstD_CA', 'comment': '', '_id': ObjectId('56cb5
1c57469f7afbc7f402f'), 'date': datetime.datetime(2013, 1, 19, 0, 0), 't
ags': ['transfer', 'transfer'], 'info': 'Transfer for Prestation Compen
satoire{FirstD‐ESA}', 'value': ‐70.0}
cursor = SAVING.find().sort([ ('date', ASCENDING) ])
print()
N=4
print("1st {} documents: ORDERED by ASCENDING date".format(N))
    print(document)

cursor = SAVING.find().sort([ ('date', DESCENDING) ])
print()
N=4
print("1st {} documents: ORDERED by DESCENDING date".format(N))
    print(document)

In [24]:
This example searches for documents with 'salary' 'salary' as tags, and projects 'value' and 'date'
fields.
This code will show the salary progression in the data over the period nice salary increase, +67% in
3.5 years ... the gravy train went through here !
{'value': ‐10.0}
{'value': ‐1200.0}
{'value': 883.65}
{'value': ‐58.4}
cursor = SAVING.find(
    # SELECTION CRITERIA:
    {},
    # PROJECTION ('columns')
    {'value': 1, '_id': 0}).
  sort([ ('date', ASCENDING) ])
print()
N=4
    print(document)

In [25]:
cursor = SAVING.find(
    {'tags': ['salary','salary']},
    {'value': 1, 'date': 1, '_id': 0}
).
  sort([ ('date', ASCENDING) ])

#print(cursor)
print(cursor.count())
print()
N=100
#wages={}
wages=[]
dates=[]
    print(document)
    # wages[ document['date'] ] = document['value']
    wages.append( document['value'] )
    dates.append( document['date'] )
    #print(document['value'])

Now let's plot that progression
44
{'date': datetime.datetime(1988, 9, 28, 0, 0), 'value': 883.65}

In [26]:
Aggregation
Now let's look at an aggregation example where we want to count the animals by type in the 'tags' field
of the data.
We first 'map' the data (using '$unwind' operation) onto a simpler structure (with only 1 tag per
document),
then we count ('reduce') the tags with the '$group/$sum' operations.
Using matplotlib backend: Qt4Agg
Populating the interactive namespace from numpy and matplotlib
%pylab
%matplotlib inline
# Note: above needs matplotlib
# conda install matplotlib
ax = plt.subplot(111)
#plt.plot(dates, wages, 'r^', label='salary progression', linestyle='solid') #drawstyle='st
ax.bar(dates, wages, width=20, label='Net salary progression [GBP]', align='center',
ax.legend(bbox_to_anchor=(1.0, 1.15)).draggable()
## Modify axis:
mybox = ax.get_position()
ax.set_position([mybox.x0, mybox.y0, mybox.width * 3.0, mybox.height*1.5])
#legend()
#?plot
#?bar

In [29]:
Now let's return to our 'Money_UK' database and use aggregation to see which categories of
expenditure were the most important by grouping on the tags whilst summing the total values.
We will then plot this as a pie chart
Out[29]:
[{'_id': 'cat', 'count': 3},
{'_id': 'dog', 'count': 2},
{'_id': 'mouse', 'count': 1}]
>>> from pymongo import MongoClient
>>> db = MongoClient().aggregation_example
>>> result = db.things.insert_many([{"x": 1, "tags": ["dog", "cat"]},
...                                 {"x": 2, "tags": ["cat"]},
...                                 {"x": 2, "tags": ["mouse", "cat", "dog"]},
...                                 {"x": 3, "tags": []}])
>>> result.inserted_ids
>>> from bson.son import SON
>>> pipeline = [
...     {"$unwind": "$tags"},
...     {"$group": {"_id": "$tags", "count": {"$sum": 1}}},
...     {"$sort": SON([("count", ‐1), ("_id", ‐1)])}
... ]
>>> list(db.things.aggregate(pipeline))

In [31]:
Opened db<Money_UK>
Available collections in db<Money_UK>: ['CARD', 'SAVING']
{'total': 12792.72, '_id': ['fun', 'hols']}
{'total': 11652.940000000002, '_id': ['misc', 'pay']}
{'total': 8805.89, '_id': ['computing', 'micro']}
{'total': 6818.120000000001, '_id': ['business', 'bus']}
{'total': 6311.000000000002, '_id': ['car', 'car']}
{'total': 4642.729999999999, '_id': ['media', 'music']}
{'total': 3439.37, '_id': ['bills', 'house']}
{'total': 2400.4499999999994, '_id': ['bills', 'clothes']}
{'total': 2184.15, '_id': ['present', 'present']}
{'total': 2053.85, '_id': ['business', 'business']}
Out[31]:
db = client.Money_UK
dbname = db.name
print("Opened db<" + dbname + ">")
collnames = [collname for collname in db.collection_names() ]
print("Available collections in db<" + dbname + ">: " + str(collnames))
from bson.son import SON
pipeline = [
     {"$group": {"_id": "$tags", "total": {"$sum": {"$abs": "$value"}}}},
     {"$sort": SON([("total", ‐1), ("_id", ‐1)])}
]
N=10
groups=[]
values=[]
for item in list(CARD.aggregate(pipeline))[:N]:
    print(item)
    groups.append (item['_id'][0]+"_"+item['_id'][1] )
    values.append (item['total'] )

pie(x=values,labels=groups)

In [32]:
Now let's perform aggregation by year to see how those categories evolved over time.
Note that we now project a new field 'year' a string derived from the 'date' field.
We then match on the year to only analyse data for that year and sort on the total as before.
Once we have aggregated the data for each year we then create a pie chart for that year, showing the
top 6 categories
([<matplotlib.patches.Wedge at 0x85f0e10>,
  <matplotlib.patches.Wedge at 0x88002e8>,
  <matplotlib.patches.Wedge at 0x8800f28>,
  <matplotlib.patches.Wedge at 0x8806ba8>,
  <matplotlib.patches.Wedge at 0x880d828>,
  <matplotlib.patches.Wedge at 0x88114a8>,
  <matplotlib.patches.Wedge at 0x8818128>,
  <matplotlib.patches.Wedge at 0x8818d68>,
  <matplotlib.patches.Wedge at 0x881e9e8>,
  <matplotlib.patches.Wedge at 0x8824668>],
[<matplotlib.text.Text at 0x87fbe48>,
  <matplotlib.text.Text at 0x8800ac8>,
  <matplotlib.text.Text at 0x8806748>,
  <matplotlib.text.Text at 0x880d3c8>,
  <matplotlib.text.Text at 0x880df28>,
  <matplotlib.text.Text at 0x8811c88>,
  <matplotlib.text.Text at 0x881e588>,
  <matplotlib.text.Text at 0x8824e48>])
%%javascript
IPython.OutputArea.auto_scroll_threshold = 9999;
// This is just to prevent scroll bars on the next section

In [33]:
1988
{'total': 923.8, '_id': ['fun', 'hols']}
{'total': 411.51, '_id': ['car', 'car']}
{'total': 361.28, '_id': ['bills', 'food']}
from bson.son import SON
def money(value):
    return int(100 * value)/100.0
N=6
for year in range(1988,2016):
    print(year)
    pipeline = [
        { "$project": {
            'year': { "$dateToString": { 'format': "%Y", 'date': "$date" } },
            'tags': 1,
            'value': 1,
            },
        },
        { "$match": { 'year': str(year) }},
        { "$group": {"_id": "$tags", "total": {"$sum": {"$abs": "$value"}}} },
        { "$sort":  SON([("total", ‐1), ("_id", ‐1)]) }
    ]
    #print( list(CARD.aggregate(pipeline))[:N] )
    groups=[]
    values=[]
    total = 0.0
    for item in list(CARD.aggregate(pipeline)):
        total += item['total']

    for item in list(CARD.aggregate(pipeline))[:N]:
        print(item)
        groups.append (item['_id'][0]+"_"+item['_id'][1] )
        values.append (item['total'] )

    pie(x=values,labels=groups)
    title = str(year) + ": " + str(money(total)) + " GBP"

    legend(title=title,labels=[])
    show()

''''''

1989
{'total': 1567.31, '_id': ['fun', 'hols']}
{'total': 1055.81, '_id': ['car', 'car']}
{'total': 1053.8000000000002, '_id': ['bills', 'clothes']}
1990
{'total': 5162.360000000001, '_id': ['fun', 'hols']}
{'total': 3035.06, '_id': ['misc', 'pay']}
{'total': 967.56, '_id': ['car', 'car']}
{'total': 819.39, '_id': ['present', 'pres']}

1991
{'total': 3856.2000000000003, '_id': ['misc', 'pay']}
{'total': 1896.34, '_id': ['fun', 'hols']}
{'total': 1048.65, '_id': ['car', 'car']}
1992
{'total': 3778.0599999999995, '_id': ['misc', 'pay']}
{'total': 2424.59, '_id': ['fun', 'hols']}
{'total': 1355.2700000000002, '_id': ['car', 'car']}
{'total': 1283.54, '_id': ['media', 'video']}
1993
{'total': 802.9300000000001, '_id': ['cashout', 'cash']}
{'total': 649.74, '_id': ['car', 'car']}
{'total': 297.0, '_id': ['misc', 'unknown']}

1994
{'total': 312.38, '_id': ['fun', 'hols']}
{'total': 49.91, '_id': ['fun', 'hotel']}
{'total': 23.33, '_id': ['car', 'car']}
{'total': 10.0, '_id': ['misc', 'pay']}
1995
{'total': 233.29, '_id': ['media', 'photo']}
{'total': 153.43, '_id': ['car', 'car']}
{'total': 119.06, '_id': ['bills', 'food']}

1996
{'total': 203.5, '_id': ['misc', 'pay']}
{'total': 132.46, '_id': ['computing', 'electronics']}
1997
{'total': 663.67, '_id': ['misc', 'pay']}
{'total': 401.40000000000003, '_id': ['car', 'car']}
{'total': 282.69, '_id': ['bills', 'dogs']}

1998
{'total': 374.73, '_id': ['bills', 'mariage']}
{'total': 370.33, '_id': ['cashout', 'cash']}
{'total': 349.02, '_id': ['computing', 'electronics']}
1999
{'total': 110.58, '_id': ['media', 'magazine']}
{'total': 104.46, '_id': ['misc', 'unknown']}
{'total': 29.39, '_id': ['media', 'books']}
{'total': 4.82, '_id': ['interest', 'int_tinc']}
{'total': 1.08, '_id': ['bills', 'telephone']}
2000
{'total': 33.81, '_id': ['misc', 'misc']}
{'total': 0.37, '_id': ['interest', 'int_tinc']}

2001
2002
{'total': 19.99, '_id': ['media', 'video']}
2003

2004
2005
2006

2007
2008
2009
{'total': 165.03, '_id': ['fun', 'hotel']}
{'total': 0.07, '_id': ['interest', 'int']}

2010
2011
2012

2013
2014
2015

Out[33]:
''

Using MongoDB and Python - Demo

Recommended

More Related Content

Viewers also liked (13)

Similar to Using MongoDB and Python - Demo (20)

Recently uploaded (20)

Using MongoDB and Python - Demo