73 lines
2.3 KiB
Python
73 lines
2.3 KiB
Python
|
#!/usr/bin/env python3
|
||
|
|
"""
|
||
|
|
数据透视汇总模板
|
||
|
|
按照指定列进行分组聚合统计
|
||
|
|
"""
|
||
|
|
|
||
|
|
import pandas as pd
|
||
|
|
import sys
|
||
|
|
from pathlib import Path
|
||
|
|
|
||
|
|
def main():
|
||
|
|
# 参数配置
|
||
|
|
file_path = "{file}"
|
||
|
|
output_path = "{output}"
|
||
|
|
group_by = "{group_by}" # 分组列（逗号分隔）
|
||
|
|
agg_column = "{agg_column}" # 要聚合的列
|
||
|
|
agg_func = "{agg_func}" # 聚合函数: sum/mean/count/max/min
|
||
|
|
|
||
|
|
# 读取文件
|
||
|
|
df = pd.read_excel(file_path)
|
||
|
|
|
||
|
|
# 解析分组列
|
||
|
|
group_columns = [col.strip() for col in group_by.split(",")]
|
||
|
|
|
||
|
|
# 检查列是否存在
|
||
|
|
missing_cols = [col for col in group_columns if col not in df.columns]
|
||
|
|
if missing_cols:
|
||
|
|
print(f"错误: 分组列不存在: {{missing_cols}}")
|
||
|
|
print(f"可用列: {{list(df.columns)}}")
|
||
|
|
sys.exit(1)
|
||
|
|
|
||
|
|
if agg_column and agg_column not in df.columns:
|
||
|
|
print(f"错误: 聚合列 '{agg_column}' 不存在")
|
||
|
|
print(f"可用列: {{list(df.columns)}}")
|
||
|
|
sys.exit(1)
|
||
|
|
|
||
|
|
# 执行透视
|
||
|
|
if agg_column:
|
||
|
|
if agg_func == "sum":
|
||
|
|
pivot = df.groupby(group_columns)[agg_column].sum().reset_index()
|
||
|
|
elif agg_func == "mean":
|
||
|
|
pivot = df.groupby(group_columns)[agg_column].mean().reset_index()
|
||
|
|
elif agg_func == "count":
|
||
|
|
pivot = df.groupby(group_columns)[agg_column].count().reset_index()
|
||
|
|
elif agg_func == "max":
|
||
|
|
pivot = df.groupby(group_columns)[agg_column].max().reset_index()
|
||
|
|
elif agg_func == "min":
|
||
|
|
pivot = df.groupby(group_columns)[agg_column].min().reset_index()
|
||
|
|
else:
|
||
|
|
pivot = df.groupby(group_columns)[agg_column].sum().reset_index()
|
||
|
|
else:
|
||
|
|
# 不指定聚合列，只计数
|
||
|
|
pivot = df.groupby(group_columns).size().reset_index(name="count")
|
||
|
|
|
||
|
|
# 保存结果
|
||
|
|
pivot.to_excel(output_path, index=False)
|
||
|
|
|
||
|
|
print(f"数据透视完成")
|
||
|
|
print(f"文件: {{file_path}}")
|
||
|
|
print(f"分组列: {{group_columns}}")
|
||
|
|
if agg_column:
|
||
|
|
print(f"聚合列: {{agg_column}}")
|
||
|
|
print(f"聚合函数: {{agg_func}}")
|
||
|
|
print(f"输出: {{output_path}}")
|
||
|
|
print(f"\n汇总行数: {{len(pivot)}}")
|
||
|
|
|
||
|
|
# 显示前几行
|
||
|
|
print("\n预览:")
|
||
|
|
print(pivot.head(10).to_string(index=False))
|
||
|
|
|
||
|
|
if __name__ == "__main__":
|
||
|
|
main()
|