HConcat of mark_geoshape和mark_bar根据顺序中断

Question

我试着用Python和牛郎星制作Vega/Vega图表。以下代码返回Javascript错误：

map_ = alt.topo_feature('https://raw.githubusercontent.com/ginseng666/GeoJSON-TopoJSON-Austria/master/2021/simplified-99.5/gemeinden_995_topo.json', 'gemeinden')

df = pd.DataFrame({
    'population': {
        10101: 3239, 10201: 1985, 10301: 1890, 10302: 1846, 10303: 2122, 10304: 3217,
        10305: 1270, 10306: 1212, 10307: 2192, 10308: 1400, 10309: 3640, 10310: 1741,
        10311: 1291, 10312: 2925, 10313: 2691, 10314: 1436, 10315: 3203, 10316: 2994,
        10317: 2083, 10318: 1717, 10319: 1974, 10320: 501,  10321: 826, 10322: 990,
        10323: 1114
    }
}).reset_index()
df.columns = ['iso', 'population']

hist = alt.Chart(df).mark_bar().encode(
    x = alt.X('population:Q', bin=True),
    y = alt.Y('count()')
)

cmap = alt.Chart(map_).mark_geoshape().encode(
    color=alt.Color('population:Q', scale=alt.Scale(scheme='spectral')),
).transform_lookup(
        lookup='properties.iso',
        from_=alt.LookupData(df, 'iso', ['population'])
)

cmap | hist

上面的代码会产生以下错误：

Javascript Error: Undefined data set name: "data_0"
This usually means there's a typo in your chart specification. See the javascript console for the full traceback.

然而，令人惊讶的是，当我反转HConcat的顺序时，显示的图表没有错误：

hist | cmap

​

​

Answer 1

有趣的发现！我相信这是下一个Vega库中的一个bug，，它似乎对数据集进行了错误的排序，以便在定义数据集之前引用一个。

解决方法是反向查找，它适用于任何一个hconcat顺序：

cmap = alt.Chart(df).mark_geoshape().encode(
    color=alt.Color('population:Q', scale=alt.Scale(scheme='spectral')),
).transform_lookup(
        lookup='iso',
        from_=alt.LookupData(map_, 'properties.iso', ['geometry', 'type'])
)

cmap | hist

​

​