如何在Plotly Express绘制的柱状图中保留2位小数？

方案1：修正texttemplate格式并全局应用 #

你尝试的texttemplate = "%{y:.2f}%"格式逻辑正确，可能是因为未确保格式应用到所有trace或文本显示位置问题，修改代码如下：

fig8 = px.bar(
    new_data,
    x=["Loyal", "Exited"],
    y=["active", "inactive"],
    title="Distribution of active membership in loyal and exited clients group",
    width=700,        
    orientation="v",
    color_discrete_map={"active": "green", "inactive": "yellow"},
    barmode="stack"
)
fig8.update_layout(
    xaxis_title_text="Types of clients", 
    yaxis_title_text="Number of clients, %", 
    barnorm="percent"
)
# 格式化文本为两位小数百分比，同时设置文本显示位置
fig8.update_traces(
    marker_line_width=0, 
    texttemplate="%{y:.2f}%",
    textposition="inside"
)
fig8.layout["legend"]["title"] = "Type of clients' membership"
fig8.show()

方案2：手动计算百分比（更可控） #

如果方案1仍不生效，建议提前在数据中计算百分比，绕过barnorm自动计算的不确定性：

# 提前计算每组的占比百分比
new_data['active_pct'] = (new_data['active'] / (new_data['active'] + new_data['inactive'])) * 100
new_data['inactive_pct'] = (new_data['inactive'] / (new_data['active'] + new_data['inactive'])) * 100

fig8 = px.bar(
    new_data,
    x=["Loyal", "Exited"],
    y=["active_pct", "inactive_pct"],
    title="Distribution of active membership in loyal and exited clients group",
    width=700,        
    orientation="v",
    color_discrete_map={"active_pct": "green", "inactive_pct": "yellow"},
    barmode="stack",
    texttemplate="%{y:.2f}%"  # 直接在绘图时设置文本格式
)
fig8.update_layout(
    xaxis_title_text="Types of clients", 
    yaxis_title_text="Number of clients, %"
)
fig8.layout["legend"]["title"] = "Type of clients' membership"
# 修正图例名称，去掉后缀
fig8.for_each_trace(lambda t: t.update(name=t.name.replace('_pct', '')))
fig8.show()

失效原因说明 #

使用barnorm="percent"时，Plotly动态计算百分比值，若trace的文本属性未正确绑定计算后的值，会导致格式化失效。方案1强制对所有trace应用格式；方案2通过提前计算让数值完全可控，避免内部计算的不确定性。

内容的提问来源于stack exchange，提问作者Victoria Yermolenko