How to dissolve overlapping polygons using GeoPandas

Question

Assuming the following:

from shapely.affinity import translate
from shapely.geometry import Polygon
import geopandas

p0 = Polygon( [(0,0), (2,0), (2,2), (0,2)] )
p1 = translate(p0, xoff=1)
p2 = translate(p0, xoff=4)

df = geopandas.GeoDataFrame(geometry=[
    p0, p1, p2
])
df.plot(cmap='cividis', alpha=0.7, edgecolor='black')

How can I get an output that dissolves polygons that overlap?

I found this answer but I am not convinced because when I replicate the answer, I get a Dataframe with 3 rows, but it turns out that the rows contain multiple Polygon objects in each row/geometry cell, so there is no real merge done.

My desired output is something like:

p3 = Polygon( [(0,0), (3,0), (3,2), (0,2), (0,0) ])
p4 = Polygon( [(4,0), (6,0), (6,2), (4,2), (4,0) ])

dfout = geopandas.GeoDataFrame(geometry=[
    p3, p4
])
dfout.plot(cmap="cividis", alpha=0.7, edgecolor="black")

Answer 1

If you are interested only in polygons (do not care about its attributes), you can use unary_union to merge all polygons and then explode resulting multipolygon.

from shapely.affinity import translate
from shapely.geometry import Polygon
import geopandas

p0 = Polygon( [(0,0), (2,0), (2,2), (0,2)] )
p1 = translate(p0, xoff=1)
p2 = translate(p0, xoff=4)

df = geopandas.GeoDataFrame(geometry=[
    p0, p1, p2
])

geoms = df.geometry.unary_union
df = geopandas.GeoDataFrame(geometry=[geoms])

df = df.explode().reset_index(drop=True)

df.plot(cmap='cividis', alpha=0.7, edgecolor='black')