Merge two GeoJSON into one in a dataframe - pandas

Question

I have a dataframe containing GeoJSON:
data = {'geojson': {0: '{"type":"LineString","coordinates":[[1,4],[2,5]]}',
1: '{"type":"LineString","coordinates":[[3,6],[4,7]]}'},
'checkpoint': {0: 6, 1: 0},'lom_name': {0: 'marathon19', 1: 'marathon19'}}
df = pd.DataFrame.from_dict(data)
The desired result is:
geojson lob_name
{"type":"LineString","coordinates":[[1,4],[2,5],[3,6],[4,7]]} marathon19
I tried df = df.groupby(['geojson']).apply(list) not really giving something i need

Answer 1

You could use the following approach using geopandas dissolve function on geojson data (a non-geometric approach would involve aggregating, concatenating and parsing the df) :
import json
import geopandas as gpd
data = { "type": "FeatureCollection",
"features": [
{ "type": "Feature",
"geometry": {
"type": "LineString",
"coordinates": [
[3,6],[4,7]
]
},
"properties": {
"checkpoint": 0,
"prop1": 'marathon19'
}
},
{ "type": "Feature",
"geometry": {
"type": "LineString",
"coordinates": [
[1,4],[2,5]
]
},
"properties": {
"checkpoint": 0,
"prop1": 'marathon19'
}
}
]
}
# create geodataframe from geojson-features
gdf = gpd.GeoDataFrame.from_features(data)
# dissolve line strings
gdf = gdf.dissolve()
# convert back to geojson
gdf.to_json()
If the input lines are non-contiguous (like in your example data) you still need to create a linestring out of the resulting multilinestring (see for example and caveats here):
import shapely
# Put the sub-line coordinates into a list of sublists
outcoords = [list(i.coords) for i in gdf.iloc()[0].geometry]
# Flatten the list of sublists and use it to make a new line
outline = shapely.geometry.LineString([i for sublist in outcoords for i in sublist])
# set geometry of gdf to linestring
gdf = gdf.set_geometry([outline])