Lineplot doesn't show all dates in axis

Question

I have the followings:

fig, ax = plt.subplots(figsize=(40, 10))
sns.lineplot(x="Date", y="KFQ imports", data=df_dry, color="BLACK", ax=ax)
sns.lineplot(x="Date", y="QRR imports", data=df_dry, color="RED",ax=ax)

ax.set(xlabel="Date", ylabel="Value", )
x_dates = df_dry['Date'].dt.strftime('%b-%Y')
ax.set_xticklabels(labels=x_dates, rotation=45)

Result

When I use a barchart (sns.barplot) the entire spectrum of dates are shown. Am I missing something for the line chart? I

Solution 1

The idea would be to set the xticks to exactly the dates in your dataframe. To this end you can use set_xticks(df.Date.values). It might then be good to use a custom formatter for the dates, which would allow to format them in the way you want them.

import pandas as pd
import matplotlib.pyplot as plt
from matplotlib import dates
import seaborn as sns

df = pd.DataFrame({"Date" : ["2018-01-22", "2018-04-04", "2018-12-06"],
                   "val"  : [1,2,3]})
df.Date = pd.to_datetime(df.Date)


ax = sns.lineplot(data=df, x="Date", y="val", marker="o")
ax.set(xticks=df.Date.values)
ax.xaxis.set_major_formatter(dates.DateFormatter("%d-%b-%Y"))
plt.show()

Note how the same can be achieved without seaborn, as

ax = df.set_index("Date").plot(x_compat=True, marker="o")
ax.set(xticks=df.Date.values)
ax.xaxis.set_major_formatter(dates.DateFormatter("%d-%b-%Y"))
plt.show()

Solution 2

One way to circumvent the date tick sparsification from Seaborn is to convert the date column to string values.

df['dates_str'] = df.dates.astype(str)

Plotting using the new column as the x-axis will show all the dates.

Comments

[ImportanceOfBeingErnest]

Just as with usual plots, you see nicely fitting labels, here, every half year. Do you want to show lables on each location that has data in the plot? (Note that in that case labels would probably overlap)

[adhg]

@ ImportanceOfBeingErnest correct, I'm aware of the overlap and this is why I added a rotation (45 or 90). I understand that the 'default' will present it neatly but for this purpose the dates (text) are critical. Thanks

[ImportanceOfBeingErnest]

So does adding xticks=df.Date into .set() do what you're after?

[adhg]

So If I understand you correctly, I did: ax.set(xlabel="Date", ylabel="mpta", ticks=df_dry['Date']) return result: AttributeError: Unknown property ticks

[adhg]

I get: Cannot compare type 'Timestamp' with type 'float'. Just to be sure: I checked and the Date is pandas.core.series.Series (type(df_dry['Date'])