Plotting column values on condition of other columns of dataframe using matplotlib against same x-axis

Question

I have a dataframe with 600 rows in the shape of :

rpttime         metric          value
25/4/2018 15:45 UTIL_CPU        1.5
25/4/2018 15:44 IDLE            74
25/4/2018 15:41 REC_BYTES_S     0
25/4/2018 15:47 ENT_CPU         100
25/4/2018 15:44 ENT_CORE        1
25/4/2018 15:48 TRANS_BYTES     92
25/4/2018 15:43 PINNED          5425
25/4/2018 15:48 PAGING_PAG      0
25/4/2018 15:48 IOPS_IN         NULL
25/4/2018 15:47 TRANS_BYTES_S   23484
25/4/2018 15:43 PAGE_OUT        0
25/4/2018 15:42 IOPS_OUT        10

I want to plot line plots for the "value" column on y-axis against "rpttime" on x-axis as time-series for each individual items in the column "metric". There are around 20 individual items in the column "metric". There are some NULL values in the "value" column which are to be omitted which is the row is to omitted for each NULL value in the "value" column. There are atleast one NULL value for each individual items of the "metric" column. All line plots for those 20 individual items are to plotted in one graph. How to approach this ?

Can you post a clean data frame with which rows with value==Null, as it's ambiguous from the posted data frame due to both NA and numerical values present? — Amey Dahale
– Amey Dahale, Commented Jun 19, 2018 at 5:47
Updated the dataframe. The NULL value appears at least once for each individual items of the "metric" column. — Arkistarvh Kltzuonstev
– Arkistarvh Kltzuonstev, Commented Jun 19, 2018 at 5:56
So do you want to ignore the rpttime where value==Null while ploting series? Or fill it with some other value? — Amey Dahale
– Amey Dahale, Commented Jun 19, 2018 at 6:00
rpttime contain string like 25/4/2018 15:48 and the dataframe has an index column starting from 0 which isn't shown here. — Arkistarvh Kltzuonstev
– Arkistarvh Kltzuonstev, Commented Jun 19, 2018 at 8:30

SpghttCd · Accepted Answer · 2018-06-19 08:21:14Z

1

If you want separate lines plotted for each metric-value, you first should groupby metric:

grpd = df.groupby('metric')

Then you can iterate over the created groups plotting each set of values against rpttime:

for name, data in grpd:
    plt.plot(data.rpttime.values, data.value.values, 'x-', label = name)
plt.legend()

Note: data you provided is not enough for a really impressive result though, as there are not multiple equals in metric:

PS: imported your data with

df = pd.read_fwf('wherever/file/may/roam', colspecs=[(None, 15), (16, 31), (32, None)])
df.rpttime = pd.to_datetime(df.rpttime)

PPS: NA is handeled automatically, i.e.: NULL-values just are not plotted (see IOPS_IN)

edited Jun 19, 2018 at 8:21

answered Jun 19, 2018 at 7:46

SpghttCd

10.9k2 gold badges23 silver badges28 bronze badges

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

Comments

Amey Dahale · Accepted Answer · 2018-06-19 06:14:31Z

0

You can directly use df.dropna() to drop Null values which aren't desired while plotting

import pandas as pd

# Original data frame
>>>df_origin

# Drop NaN values using dropna
df_to_be_plotted = df_origin.dropna()
>>> df_to_be_plotted

answered Jun 19, 2018 at 6:14

Amey Dahale

7506 silver badges10 bronze badges

2 Comments

Arkistarvh Kltzuonstev Over a year ago

Dropping NULL values is ok, any solution on plotting multiple time-series data of column "value" with individual items of "metric" column.

Amey Dahale Over a year ago

I guess you can take a view from original data frame for each unique item in metric column to plot a time series against values in value column. If you can elaborate more on plotting part in question it'll be helpful

Collectives™ on Stack Overflow

Plotting column values on condition of other columns of dataframe using matplotlib against same x-axis

2 Answers 2

Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related