Read Dataframes

The central command to read Pandas dataframes is the function “read”, see the Official Reference on read .

pd.read_<file_format>

Though different file formats are supported, here only csv and Excel formats are described.

Read a dataframe from a csv file

If the columns of your data are not delimited by a comma, but by an other sign, e.g. a semi-colon, then you need to specify this delimiter.

df = pd.read_csv(fullPathName, delimiter=";")

Read a dataframe from an Excel file

Read with conditions on columns

You can restrict the columns imported to those that you really want to use.

colsIwannaUse = ["Name", "Cylinders"]
df = pd.read_csv(fullPathName, usecols = colsIwannaUse)

Alteratively, you can specify the positions of the columns.

colsIwannaUse = fullPathName = stem +  "/ZZ_Data/co2.csv"[0, 2, 6]
df = pd.read_csv(fullPathName, usecols = colsIwannaUse)

Read with conditions on rows

You can specify the row in which the column names (or the ‘header’) is positioned, if they are not in the uppermost row.
Further, you can limit the number of rows to be imported to the dataframe with the parameter ‘nrows’.

df = pd.read_csv(fullPathName, header=17, nrows=4)

You can skip the first n rows or the last m rows

df = pd.read_csv(fullPathName, skiprows=10, skipfooter=11)

Certainly, you can combine most of these parameters, though not all of these combination make sense.

Read and change values while reading

For the cars dataset, you might want the number of cylinders to be an integer.

df = pd.read_csv(fullPathName, converters={"Cylinders": int()})

You can define so-called converters, i.e. functions that you include to the read function.

def square(x):
    x = int(x)
    return x*x
def extract(x):
    return x[0:2]

Call them with the syntax of the converter.

df = pd.read_csv(fullPathName, converters={"Cylinders": square, "Horsepower": extract})
df.head(3)

Well, a bit meaningless here, just for the sake of demonstration.