I want to transform a table from a csv into a hierarchy. The table has the following format:
lvl1 lvl2 lvl3 size
A a 1 1
A a 2 2
A b 1 3
A b 2 4
B a 1 5
B a 2 6
B b 1 7
B b 2 8
Which translates to the JSON object below.
I used d3.nest like this:
var n = d3.nest()
.key(d => d.lvl1)
.key(d => d.lvl2)
.key(d => d.lvl3)
.entries(d)
which results in the following format which is almost what I want to pass to d3.hierarchy:
[{
"key": "A",
"values": [{
"key": "a",
"values": [{
"key": "1",
"values": [{
"lvl1": "A",
"lvl2": "a",
"lvl3": 1,
"size": 1
}]
},
{
"key": "2",
"values": [{
"lvl1": "A",
"lvl2": "a",
"lvl3": 2,
"size": 2
}]
}
]
}]
}
//...
]
However, this adds one level too many and ideally I would like to have the following output:
[{
"key": "A",
"values": [{
"key": "a",
"values": [{
"key": "1",
"size": 1
},
{
"key": "2",
"size": 2
}
]
}]
} //,...
]
This format I could conveniently pass down to d3.hierarchy(n, d => d.values). How to achieve that?
N.B. I don't care if there are additional fields in the leaf, thus, an output like this would also be fine as long as the leaf is not nested again:
[{
"key": "A",
"values": [{
"key": "a",
"values": [{
"key": "1",
"lvl1": "A",
"lvl2": "a",
"lvl3": 1,
"size": 1
},
{
"key": "2",
"lvl1": "A",
"lvl2": "a",
"lvl3": 2,
"size": 2
}
]
}]
} //,...
]
JSON
[{
"lvl1": "A",
"lvl2": "a",
"lvl3": 1,
"size": 1
},
{
"lvl1": "A",
"lvl2": "a",
"lvl3": 2,
"size": 2
},
{
"lvl1": "A",
"lvl2": "b",
"lvl3": 1,
"size": 3
},
{
"lvl1": "A",
"lvl2": "b",
"lvl3": 2,
"size": 4
},
{
"lvl1": "B",
"lvl2": "a",
"lvl3": 1,
"size": 5
},
{
"lvl1": "B",
"lvl2": "a",
"lvl3": 2,
"size": 6
},
{
"lvl1": "B",
"lvl2": "b",
"lvl3": 1,
"size": 7
},
{
"lvl1": "B",
"lvl2": "b",
"lvl3": 2,
"size": 8
}
]