Docker file to run Python with Pandas

Question

I have created the following Dockerfile

FROM python
COPY . /home
CMD pip install pandas
CMD mkdir /home/report
CMD mkdir /home/data
CMD python /home/hello.py

where hello.py is the simple Python script

name = input('What is your Name? ')
print('Nice to meet you', name)

from pandas import read_csv
mydf = read_csv('mycsv.csv')
print(mydf.head())

I then build the Docker image with docker build -t myexample . and run it with docker run -it myexample bash so as to interact with it via the shell. The building goes fine and upon running it I presented with the shell prompt, but then:

No directories report or data have been created under /home.
The last command python /home/hello.py does not execute on its own. I have to type it myself to get the script to run.
Once I type python /home/hello.py, the first two lines that greet and prompt for my name are executed properly, but then an error says that pandas is unknown.
It is not until I install pandas manually that the whole script runs correctly.

So, in summary, it seems that none of the CMD statements were taken into account. What am I doing wrong?

Only the last CMD will be valid, you should replace all other CMD by one RUN — user2915097
– user2915097, Commented Jun 3, 2017 at 17:37

Robert · Accepted Answer · 2017-06-04 02:38:25Z

4

When you build an image use RUN to execute commands. Then, use CMD only once to declare the command that will start the container after the build (so there is only CMD):

Dockerfile:

FROM python
RUN pip install pandas
RUN mkdir /home/report
RUN mkdir /home/data

COPY . /home
WORKDIR /home
CMD python /home/hello.py

mycsv.csv:

a,b,c
1,2,3
4,5,6

Build with: docker build . -t pandas-test

Run:

▶ docker run -it pandas-test
What is your Name? Tfovid
Nice to meet you Tfovid
   a  b  c
0  1  2  3
1  4  5  6

answered Jun 4, 2017 at 2:38

Robert

37.6k8 gold badges97 silver badges101 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Ricardo Branco · Accepted Answer · 2017-06-04 02:56:02Z

3

FROM python
RUN pip install --no-cache-dir pandas && \
    mkdir /home/report /home/data && \
    chmod +x /home/hello.py
COPY . /home
VOLUME /home/report /home/data
WORKDIR /home
ENTRYPOINT /home/hello.py

Notes:

As a general rule, it's better to collapse 2 RUN statements into one so one layer is created.
I suggest you use a better directory than /home
Use volumes to store data and logs.

answered Jun 4, 2017 at 2:56

Ricardo Branco

6,2191 gold badge26 silver badges33 bronze badges

1 Comment

Tfovid Over a year ago

Why do you suggest to use another directory than /home? Could you please give an example?

Collectives™ on Stack Overflow

Docker file to run Python with Pandas

2 Answers 2

Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related