Skip to main content

Fill NA values

Use this function to fill in NA values in a sequential manner up to a specified limit. When using this function, you will specify the following:

  • Whether the method to fill the NA values should go forward (default) or backward.
  • Whether the NA values should be filled along rows (default) or columns.
  • The maximum number of consecutive NA values to fill (defaults to 1).
import h2o
h2o.init()

# Create a random data frame with 10 rows and 3 columns. 
# Specify that no more than 20% of the values are NAs.
df = h2o.create_frame(rows=10, 
                      cols=3, 
                      real_fraction=1.0, 
                      real_range=100, 
                      missing_fraction=0.2, 
                      seed=123)
df
      C1        C2        C3
--------  --------  --------
nan       nan       -77.1047
-93.6409  -13.6593   57.4439
-93.71     25.4342   39.1013
-95.8291  -92.4271   55.4314
 84.6372  -43.4759   53.1715
-57.9583   27.4148  -26.9013
 83.0921  -62.7819  -91.9426
-77.9814   64.3228  -93.954
nan       -80.6142  nan
 27.1672   60.5492  -13.2275

[10 rows x 3 columns]

# Forward fill a row. In Python, the values for axis are 0 (row-wise) and 1 (column-wise)
filled = df.fillna(method="forward",axis=0,maxlen=1)
filled

filled
      C1        C2        C3
--------  --------  --------
nan       nan       -77.1047
-93.6409  -13.6593   57.4439
-93.71     25.4342   39.1013
-95.8291  -92.4271   55.4314
 84.6372  -43.4759   53.1715
-57.9583   27.4148  -26.9013
 83.0921  -62.7819  -91.9426
-77.9814   64.3228  -93.954
-77.9814  -80.6142  -93.954
27.1672   60.5492  -13.2275

[10 rows x 3 columns]
Feedback