Vega串联图表跨图筛选失效：单数据源下点击左图无法过滤右图

我瞅了下你的Vega-Lite配置，问题出在selection的字段匹配逻辑上。你的左侧图表用了count聚合，生成的是每个Major_Genre的统计数据，但默认的single selection会选中整个聚合后的数据对象（包含Major_Genre和聚合后的count值）。而右侧图表基于原始数据，每条数据有Title、Major_Genre、Country_Origin这些字段，和选中的聚合数据点结构不匹配，导致过滤逻辑找不到对应条目，自然就没效果了。

修复方案 #

只需要给左侧的selection明确指定要捕获的字段——也就是我们用来过滤的Major_Genre，这样选中的就是具体的类型值，而非整个数据对象，右侧的filter就能正确匹配原始数据了。

修改后的完整配置 #

{
  "config": {"view": {"width": 400, "height": 300}},
  "hconcat": [
    {
      "width": 100,
      "selection": {
        "SELECT": {
          "type": "single",
          "resolve": "global",
          "fields": ["Major_Genre"] // 新增：明确指定要捕获的字段
        }
      },
      "data": {"name": "table_data"},
      "mark": "bar",
      "encoding": {
        "x": {"type": "nominal", "field": "Major_Genre"},
        "y": {"aggregate": "count", "type": "quantitative"},
        "color": {"type": "nominal", "field": "Major_Genre"},
        "fillOpacity": {
          "condition": {"selection": "SELECT", "value": 1},
          "value": 0.3
        }
      }
    },
    {
      "width": 100,
      "data": {"name": "table_data"},
      "transform": [{"filter": {"selection": "SELECT"}}],
      "mark": "bar",
      "encoding": {
        "x": {"type": "nominal", "field": "Major_Genre"},
        "y": {"aggregate": "count", "type": "quantitative"},
        "color": {"type": "nominal", "field": "Major_Genre"}
      }
    }
  ],
  "datasets": {
    "table_data": [
      {"Title": "Cidade de Deus", "Major_Genre": "Drama", "Country_Origin": "Brazil"},
      {"Title": "Chocolate: Deep Dark Secrets", "Major_Genre": "Thriller/Suspense", "Country_Origin": "India"},
      {"Title": "Fiza", "Major_Genre": "Drama", "Country_Origin": "India"},
      {"Title": "First Love, Last Rites", "Major_Genre": "Drama", "Country_Origin": "United States"},
      {"Title": "Foolish", "Major_Genre": "Comedy", "Country_Origin": "United States"},
      {"Title": "I Married a Strange Person", "Major_Genre": "Comedy", "Country_Origin": "United States"}
    ]
  }
}

关键修改说明 #

• 在左侧的SELECT selection中添加了"fields": ["Major_Genre"]，告诉Vega我们只需要捕获这个字段的具体值，点击左侧条形时，selection会记录对应的类型名称。
• 右侧的filter逻辑不需要修改，现在它能通过SELECT捕获的Major_Genre值，正确过滤原始数据中对应的条目。

这样调整后，点击左侧任意条形，右侧图表就会自动过滤并显示对应类型的计数了。

内容的提问来源于stack exchange，提问作者Mattijn