You can run this notebook in a live session Binder or view it on Github.

DataFrames: Read and Write Data

Dask Dataframes can read and store data in many of the same formats as Pandas dataframes. In this example we read and write data with the popular CSV and Parquet formats, and discuss best practices when using these formats.

[1]:
from IPython.display import HTML

HTML('<iframe width="560" height="315" src="https://www.youtube.com/embed/0eEsIA0O1iE?rel=0&amp;controls=0&amp;showinfo=0" frameborder="0" allowfullscreen></iframe>')
/home/travis/miniconda/envs/test/lib/python3.7/site-packages/IPython/core/display.py:694: UserWarning: Consider using IPython.display.IFrame instead
  warnings.warn("Consider using IPython.display.IFrame instead")
[1]:

Start Dask Client for Dashboard

Starting the Dask Client is optional. It will provide a dashboard which is useful to gain insight on the computation.

The link to the dashboard will become visible when you create the client below. We recommend having it open on one side of your screen while using your notebook on the other side. This can take some effort to arrange your windows, but seeing them both at the same is very useful when learning.

[2]:
from dask.distributed import Client
client = Client(n_workers=1, threads_per_worker=4, processes=False, memory_limit='2GB')
client
[2]:

Client

Cluster

  • Workers: 1
  • Cores: 4
  • Memory: 2.00 GB

Create artificial dataset

First we create an artificial dataset and write it to many CSV files.

You don’t need to understand this section, we’re just creating a dataset for the rest of the notebook.

[3]:
import dask
df = dask.datasets.timeseries()
df
[3]:
Dask DataFrame Structure:
id name x y
npartitions=30
2000-01-01 int64 object float64 float64
2000-01-02 ... ... ... ...
... ... ... ... ...
2000-01-30 ... ... ... ...
2000-01-31 ... ... ... ...
Dask Name: make-timeseries, 30 tasks
[4]:
import os
import datetime

if not os.path.exists('data'):
    os.mkdir('data')

def name(i):
    """ Provide date for filename given index

    Examples
    --------
    >>> name(0)
    '2000-01-01'
    >>> name(10)
    '2000-01-11'
    """
    return str(datetime.date(2000, 1, 1) + i * datetime.timedelta(days=1))

df.to_csv('data/*.csv', name_function=name);

Read CSV files

We now have many CSV files in our data directory, one for each day in the month of January 2000. Each CSV file holds timeseries data for that day. We can read all of them as one logical dataframe using the dd.read_csv function with a glob string.

[5]:
!ls data/*.csv | head
data/2000-01-01.csv
data/2000-01-02.csv
data/2000-01-03.csv
data/2000-01-04.csv
data/2000-01-05.csv
data/2000-01-06.csv
data/2000-01-07.csv
data/2000-01-08.csv
data/2000-01-09.csv
data/2000-01-10.csv
[6]:
!head data/2000-01-01.csv
timestamp,id,name,x,y
2000-01-01 00:00:00,1047,Hannah,0.4953541470475098,0.8238577255373991
2000-01-01 00:00:01,969,Michael,0.3250317494598991,0.6590353389513288
2000-01-01 00:00:02,1018,Norbert,-0.7113490977378836,0.5976626641387559
2000-01-01 00:00:03,934,Ray,-0.865254983847219,-0.09289843564806799
2000-01-01 00:00:04,969,Tim,-0.7355338140085299,0.25423945562393135
2000-01-01 00:00:05,1023,Xavier,-0.006950890892156947,-0.7337350574035115
2000-01-01 00:00:06,994,George,0.6608749237340636,0.9994063883236972
2000-01-01 00:00:07,1005,Ursula,-0.5313278364344296,0.5107691379411325
2000-01-01 00:00:08,996,George,0.7352458285000505,-0.6898406001059514
[7]:
!head data/2000-01-30.csv
timestamp,id,name,x,y
2000-01-30 00:00:00,1017,Patricia,-0.13261279755233724,0.4876536975055663
2000-01-30 00:00:01,979,Sarah,0.8750296689968415,-0.6174189512729937
2000-01-30 00:00:02,1022,Ray,-0.9677546956635685,-0.46196125854540093
2000-01-30 00:00:03,1011,Quinn,-0.951608713585544,0.5699522290894143
2000-01-30 00:00:04,1025,Wendy,-0.40124662334730354,-0.639294238221445
2000-01-30 00:00:05,1026,Wendy,0.3904302204759287,0.9493702154354633
2000-01-30 00:00:06,1097,Wendy,0.5512923310561206,-0.3734803295633853
2000-01-30 00:00:07,933,Alice,0.08599847870869004,-0.1108551431370215
2000-01-30 00:00:08,997,Charlie,-0.7511670286347543,-0.3127062533269003

We can read one file with pandas.read_csv or many files with dask.dataframe.read_csv

[8]:
import pandas as pd

df = pd.read_csv('data/2000-01-01.csv')
df.head()
[8]:
timestamp id name x y
0 2000-01-01 00:00:00 1047 Hannah 0.495354 0.823858
1 2000-01-01 00:00:01 969 Michael 0.325032 0.659035
2 2000-01-01 00:00:02 1018 Norbert -0.711349 0.597663
3 2000-01-01 00:00:03 934 Ray -0.865255 -0.092898
4 2000-01-01 00:00:04 969 Tim -0.735534 0.254239
[9]:
import dask.dataframe as dd

df = dd.read_csv('data/2000-*-*.csv')
df
[9]:
Dask DataFrame Structure:
timestamp id name x y
npartitions=30
object int64 object float64 float64
... ... ... ... ...
... ... ... ... ... ...
... ... ... ... ...
... ... ... ... ...
Dask Name: from-delayed, 90 tasks
[10]:
df.head()
[10]:
timestamp id name x y
0 2000-01-01 00:00:00 1047 Hannah 0.495354 0.823858
1 2000-01-01 00:00:01 969 Michael 0.325032 0.659035
2 2000-01-01 00:00:02 1018 Norbert -0.711349 0.597663
3 2000-01-01 00:00:03 934 Ray -0.865255 -0.092898
4 2000-01-01 00:00:04 969 Tim -0.735534 0.254239

Tuning read_csv

The Pandas read_csv function has many options to help you parse files. The Dask version uses the Pandas function internally, and so supports many of the same options. You can use the ? operator to see the full documentation string.

[11]:
pd.read_csv?
[12]:
dd.read_csv?

In this case we use the parse_dates keyword to parse the timestamp column to be a datetime. This will make things more efficient in the future. Notice that the dtype of the timestamp column has changed from object to datetime64[ns].

[13]:
df = dd.read_csv('data/2000-*-*.csv', parse_dates=['timestamp'])
df
[13]:
Dask DataFrame Structure:
timestamp id name x y
npartitions=30
datetime64[ns] int64 object float64 float64
... ... ... ... ...
... ... ... ... ... ...
... ... ... ... ...
... ... ... ... ...
Dask Name: from-delayed, 90 tasks

Do a simple computation

Whenever we operate on our dataframe we read through all of our CSV data so that we don’t fill up RAM. This is very efficient for memory use, but reading through all of the CSV files every time can be slow.

[14]:
%time df.groupby('name').x.mean().compute()
CPU times: user 5.37 s, sys: 695 ms, total: 6.06 s
Wall time: 4.1 s
[14]:
name
Alice      -0.002328
Bob         0.000385
Charlie     0.000251
Dan         0.000456
Edith       0.004285
Frank      -0.002597
George     -0.002654
Hannah     -0.002461
Ingrid     -0.000397
Jerry       0.001155
Kevin       0.000827
Laura      -0.000138
Michael    -0.001992
Norbert     0.003082
Oliver     -0.000715
Patricia    0.001206
Quinn      -0.002850
Ray         0.002874
Sarah       0.002928
Tim        -0.000412
Ursula     -0.000843
Victor     -0.003662
Wendy       0.003373
Xavier      0.000907
Yvonne      0.001188
Zelda       0.001108
Name: x, dtype: float64
[ ]:

Write to Parquet

Instead, we’ll store our data in Parquet, a format that is more efficient for computers to read and write.

[15]:
df.to_parquet('data/2000-01.parquet', engine='pyarrow')
[16]:
!ls data/2000-01.parquet/
_common_metadata  part.15.parquet  part.22.parquet  part.2.parquet
_metadata         part.16.parquet  part.23.parquet  part.3.parquet
part.0.parquet    part.17.parquet  part.24.parquet  part.4.parquet
part.10.parquet   part.18.parquet  part.25.parquet  part.5.parquet
part.11.parquet   part.19.parquet  part.26.parquet  part.6.parquet
part.12.parquet   part.1.parquet   part.27.parquet  part.7.parquet
part.13.parquet   part.20.parquet  part.28.parquet  part.8.parquet
part.14.parquet   part.21.parquet  part.29.parquet  part.9.parquet

Read from Parquet

[17]:
df = dd.read_parquet('data/2000-01.parquet', engine='pyarrow')
df
[17]:
Dask DataFrame Structure:
timestamp id name x y
npartitions=30
datetime64[ns] int64 object float64 float64
... ... ... ... ...
... ... ... ... ... ...
... ... ... ... ...
... ... ... ... ...
Dask Name: read-parquet, 30 tasks
[18]:
%time df.groupby('name').x.mean().compute()
CPU times: user 1.64 s, sys: 114 ms, total: 1.75 s
Wall time: 1.21 s
[18]:
name
Alice      -0.002328
Bob         0.000385
Charlie     0.000251
Dan         0.000456
Edith       0.004285
Frank      -0.002597
George     -0.002654
Hannah     -0.002461
Ingrid     -0.000397
Jerry       0.001155
Kevin       0.000827
Laura      -0.000138
Michael    -0.001992
Norbert     0.003082
Oliver     -0.000715
Patricia    0.001206
Quinn      -0.002850
Ray         0.002874
Sarah       0.002928
Tim        -0.000412
Ursula     -0.000843
Victor     -0.003662
Wendy       0.003373
Xavier      0.000907
Yvonne      0.001188
Zelda       0.001108
Name: x, dtype: float64

Select only the columns that you plan to use

Parquet is a column-store, which means that it can efficiently pull out only a few columns from your dataset. This is good because it helps to avoid unnecessary data loading.

[19]:
%%time
df = dd.read_parquet('data/2000-01.parquet', columns=['name', 'x'], engine='pyarrow')
df.groupby('name').x.mean().compute()
CPU times: user 1.45 s, sys: 88 ms, total: 1.54 s
Wall time: 1.11 s
[19]:
name
Alice      -0.002328
Bob         0.000385
Charlie     0.000251
Dan         0.000456
Edith       0.004285
Frank      -0.002597
George     -0.002654
Hannah     -0.002461
Ingrid     -0.000397
Jerry       0.001155
Kevin       0.000827
Laura      -0.000138
Michael    -0.001992
Norbert     0.003082
Oliver     -0.000715
Patricia    0.001206
Quinn      -0.002850
Ray         0.002874
Sarah       0.002928
Tim        -0.000412
Ursula     -0.000843
Victor     -0.003662
Wendy       0.003373
Xavier      0.000907
Yvonne      0.001188
Zelda       0.001108
Name: x, dtype: float64

Here the difference is not that large, but with larger datasets this can save a great deal of time.