/Mini-Project-2-NoSQL-Databases

Mini-Project-2-NoSQL-Databases

1.preparing a large data source and loading it into both databases

We have found a datasheet containing 36634 insurence samples of growth. Datasheet Download, click here...

Neo4j populations code

LOAD CSV WITH HEADERS FROM "file:///FL_insurance_sample.csv" 
AS row MERGE (Insurance:Insurance {id:row.policyID}) 
SET 
Insurance.statecode = row.statecode, 
Insurance.county = row.county, 
Insurance.row.eq_site_limit = toFloat(row.eq_site_limit), 
Insurance.hu_site_limit = toFloat(row.hu_site_limit), 
Insurance.fl_site_limit = toFloat(row.fl_site_limit), 
Insurance.fr_site_limit = toFloat(row.fr_site_limit), 
Insurance.tiv_2011 = toFloat(row.tiv_2011), 
Insurance.tiv_2012 = toFloat(row.tiv_2012), 
Insurance.eq_site_deductible = toFloat(row.eq_site_deductible), 
Insurance.hu_site_deductible = toFloat(row.hu_site_deductible), 
Insurance.fl_site_deductible = toFloat(row.fl_site_deductible), 
Insurance.fr_site_deductible = toFloat(row.fr_site_deductible), 
Insurance.point_latitude = toFloat(row.point_latitude), 
Insurance.point_longitude = toFloat(row.point_longitude), 
Insurance.line = row.line, 
Insurance.construction = row.construction,
Insurance.point_granularity= toInteger(row.point_granularity),

MongoDB

  1. Cd path to mongofile
  2. whrite "mongod"
  3. whrite "mongo"

Now the server is startet and ready to use Test by running the command in MongoDB terminal

show dbs

now load csv file to mongodb (only posibble from cmd) 5. cd mongo path folder and drop the csv-file in here 6. run the command below:

```
mongoimport -d insurence -c insurence --type csv --file recordsinsurence.csv --headerline
```
  1. check in the mongo terminal: "show dbs" there shoould now be a now database named "insurence"

2. selecting relevant database operations, which can be used to compare the databases

1.Population time

  • MongoDB population time 596ms alt text
  • Neo4j population time 409011ms alt text

2. Insert single elements The code below show how we can insert one element of data i both database

Neo4j

CREATE (Insurance:Insurance)
SET 
Insurance.id = 172534,
Insurance.statecode = 123456789, 
Insurance.county = "CLAY COUNTY",
Insurance.row.eq_site_limit = 2,
Insurance.hu_site_limit = 254281.5,
Insurance.fl_site_limit = 246144.49,
Insurance.fr_site_limit = 0,
Insurance.tiv_2011 = 0,
Insurance.tiv_2012 = 0,
Insurance.eq_site_deductible = 123456789,
Insurance.hu_site_deductible = 123456789,
Insurance.fl_site_deductible = 123456789,
Insurance.fr_site_deductible = 123456789,
Insurance.point_latitude = 123456789,
Insurance.point_longitude = 123456789,
Insurance.line = "Residential",
Insurance.construction = "Wood",
Insurance.point_granularity = 1

MongoDB

db.insurence.insertOne({
 "_id" : ObjectId("5ebb020e2050d9540fc2d08c"), "policyID" : 172534, "statecode" : "FL", 
 "county" : "CLAY COUNTY", "eq_site_limit" : 0, "hu_site_limit" : 254281.5, "fl_site_limit" : 0,
 "fr_site_limit" : 254281.5, "tiv_2011" : 254281.5, "tiv_2012" : 246144.49, "eq_site_deductible" : 0,
 "hu_site_deductible" : 0, "fl_site_deductible" : 0, "fr_site_deductible" : 0, "point_latitude" : 30.060614, 
 "point_longitude" : -81.702675, "line" : "Residential", "construction" : "Wood", "point_granularity" : 1 
})

3. How much space does the two database consume

Neo4j

  • Before = 2.12 GB (2.278.607.182 bytes)
  • After = 2.37 GB (2.547.756.160 bytes)
  • Storage taken = 0.25 GB (269.148.978 bytes)

MongoDB

  • List item

3. Selecet all elements from db

Neo4j

Match(n) RETURN n limit(36634)

Result was 0.798 Seconds

MongoDB

DBQuery.shellBatchSize = 36634
db.insurence.find() - query all elments Result 10.00 sec

DBQuery.shellBatchSize - makes mongoDB retunring 36634 element instead of its standard db.insurence.find() - quries all elements in the insurence collection

Result was 10.00 seconds

4. How easy is it to setup

Neo4j very intuitive and easy to use GUI with alot of help, with error syntax and error handling.

MongoDB Experience with a command prompt needed, no GUI to visualize data.

3. reporting the results and conclusions

Based on our experiments with Neo4j and MongoDB can we conclude that if you have a very large data collection you will have to have alot space available since Neo4j takes up approximately 83 times as much space. But on the other hand its more intuitive to use Neo4j since you get alot of help with errors and syntax. personally we found it easier to import data into Neo4j contra MongoDB.

  • List item
  • List item

Futher work Scaling? how can they work together? more in depth with pros and cons?