最近在做报表统计方面的需求，涉及到行转列报表。根据以往经验使用SQL可以比较容易完成，这次决定挑战一下直接通过代码方式完成行转列。期间遇到几个问题和用到的新知识这里整理记录一下。
阅读目录
问题介绍
以家庭月度费用为例，可以在[Name,Area,Month]三个维度上随意组合进行分组，三个维度中选择一个做为列显示。
/// <summary> /// 家庭费用情况 /// </summary> public class House { /// <summary> /// 户主姓名 /// </summary> public string Name { get; set; } /// <summary> /// 所属行政区域 /// </summary> public string Area { get; set; } /// <summary> /// 月份 /// </summary> public string Month { get; set; } /// <summary> /// 电费金额 /// </summary> public double DfMoney { get; set; } /// <summary> /// 水费金额 /// </summary> public double SfMoney { get; set; } /// <summary> /// 燃气金额 /// </summary> public double RqfMoney { get; set; } }
| 户主-月明细报表 | ||||||
| 户主姓名 | 2016-01 | 2016-02 | ||||
|---|---|---|---|---|---|---|
| 电费 | 水费 | 燃气费 | 电费 | 水费 | 燃气费 | |
| 张三 | 240.9 | 30 | 25 | 167 | 24.5 | 17.9 |
| 李四 | 56.7 | 24.7 | 13.2 | 65.2 | 18.9 | 14.9 |
| 区域-月明细报表 | ||||||
| 区域 | 2016-01 | 2016-02 | ||||
|---|---|---|---|---|---|---|
| 电费 | 水费 | 燃气费 | 电费 | 水费 | 燃气费 | |
| 江夏区 | 2240.9 | 330 | 425 | 5167 | 264.5 | 177.9 |
| 洪山区 | 576.7 | 264.7 | 173.2 | 665.2 | 108.9 | 184.9 |
| 区域月份-户明细报表 | |||||||
| 区域 | 月份 | 张三 | 李四 | ||||
|---|---|---|---|---|---|---|---|
| 燃气费 | 电费 | 水费 | 燃气费 | 电费 | 水费 | ||
| 江夏区 | 2016-01 | 2240.9 | 330 | 425 | 5167 | 264.5 | 177.9 |
| 洪山区 | 2016-01 | 576.7 | 264.7 | 173.2 | 665.2 | 108.9 | 184.9 |
| 江夏区 | 2016-02 | 3240.9 | 430 | 525 | 6167 | 364.5 | 277.9 |
| 洪山区 | 2016-02 | 676.7 | 364.7 | 273.2 | 765.2 | 208.9 | 284.9 |
现在后台查出来的数据是List<House>类型，前台传过来分组维度和动态列字段。
第1个表格前台传给后台参数
{DimensionList:['Name'],DynamicColumn:'Month'}
第2个表格前台传给后台参数
{DimensionList:['Area'],DynamicColumn:'Month'}
第3个表格前台传给后台参数
{DimensionList:['Area','Month'],DynamicColumn:'Name'}
问题描述清楚后，仔细分析后你就会发现这里的难题在于动态分组，也就是怎么根据前台传过来的多个维度对List进行分组。
动态Linq
下面使用System.Linq.Dynamic完成行转列功能，Nuget上搜索System.Linq.Dynamic即可下载该包。
代码进行了封装，实现了通用的List<T>行转列功能。
/// <summary> /// 动态Linq方式实现行转列 /// </summary> /// <param name="list">数据</param> /// <param name="DimensionList">维度列</param> /// <param name="DynamicColumn">动态列</param> /// <returns>行转列后数据</returns> private static List<dynamic> DynamicLinq<T>(List<T> list, List<string> DimensionList, string DynamicColumn, out List<string> AllDynamicColumn) where T : class { //获取所有动态列 var columnGroup = list.GroupBy(DynamicColumn, "new(it as Vm)") as IEnumerable<IGrouping<dynamic, dynamic>>; List<string> AllColumnList = new List<string>(); foreach (var item in columnGroup) { if (!string.IsNullOrEmpty(item.Key)) { AllColumnList.Add(item.Key); } } AllDynamicColumn = AllColumnList; var dictFunc = new Dictionary<string, Func<T, bool>>(); foreach (var column in AllColumnList) { var func = DynamicExpression.ParseLambda<T, bool>(string.Format("{0}=="{1}"", DynamicColumn, column)).Compile(); dictFunc[column] = func; } //获取实体所有属性 Dictionary<string, PropertyInfo> PropertyInfoDict = new Dictionary<string, PropertyInfo>(); Type type = typeof(T); var propertyInfos = type.GetProperties(BindingFlags.Instance | BindingFlags.Public); //数值列 List<string> AllNumberField = new List<string>(); foreach (var item in propertyInfos) { PropertyInfoDict[item.Name] = item; if (item.PropertyType == typeof(int) || item.PropertyType == typeof(double) || item.PropertyType == typeof(float)) { AllNumberField.Add(item.Name); } } //分组 var dataGroup = list.GroupBy(string.Format("new ({0})", string.Join(",", DimensionList)), "new(it as Vm)") as IEnumerable<IGrouping<dynamic, dynamic>>; List<dynamic> listResult = new List<dynamic>(); IDictionary<string, object> itemObj = null; T vm2 = default(T); foreach (var group in dataGroup) { itemObj = new ExpandoObject(); var listVm = group.Select(e => e.Vm as T).ToList(); //维度列赋值 vm2 = listVm.FirstOrDefault(); foreach (var key in DimensionList) { itemObj[key] = PropertyInfoDict[key].GetValue(vm2); } foreach (var column in AllColumnList) { vm2 = listVm.FirstOrDefault(dictFunc[column]); if (vm2 != null) { foreach (string name in AllNumberField) { itemObj[name + column] = PropertyInfoDict[name].GetValue(vm2); } } } listResult.Add(itemObj); } return listResult; }
标红部分使用了System.Linq.Dynamic动态分组功能，传入字符串即可分组。使用了dynamic类型，关于dynamic介绍可以参考其它文章介绍哦。
System.Linq.Dynamic其它用法
上面行转列代码见识了System.Linq.Dynamic的强大，下面再介绍一下会在开发中用到的方法。
Where过滤
list.Where("Name=@0", "张三")
上面用到了参数化查询，实现了查找姓名是张三的数据，通过这段代码你或许感受不到它的好处。但是和EntityFramework结合起来就可以实现动态拼接SQL的功能了。
/// <summary> /// EF实体查询封装 /// </summary> /// <typeparam name="T">实体类型</typeparam> /// <param name="Query">IQueryable对象</param> /// <param name="gridParam">过滤条件</param> /// <returns>查询结果</returns> public static EFPaginationResult<T> PageQuery<T>(this IQueryable<T> Query, QueryCondition gridParam) { //查询条件 EFFilter filter = GetParameterSQL<T>(gridParam); var query = Query.Where(filter.Filter, filter.ListArgs.ToArray()); //查询结果 EFPaginationResult<T> result = new EFPaginationResult<T>(); if (gridParam.IsPagination) { int PageSize = gridParam.PageSize; int PageIndex = gridParam.PageIndex < 0 ? 0 : gridParam.PageIndex; //获取排序信息 string sort = GetSort(gridParam, typeof(T).FullName); result.Data = query.OrderBy(sort).Skip(PageIndex * PageSize).Take(PageSize).ToList<T>(); if (gridParam.IsCalcTotal) { result.Total = query.Count(); result.TotalPage = Convert.ToInt32(Math.Ceiling(result.Total * 1.0 / PageSize)); } else { result.Total = result.Data.Count(); } } else { result.Data = query.ToList(); result.Total = result.Data.Count(); } return result; }
/// <summary> /// 通过查询条件,获取参数化查询SQL /// </summary> /// <param name="gridParam">过滤条件</param> /// <returns>过滤条件字符</returns> private static EFFilter GetParameterSQL<T>(QueryCondition gridParam) { EFFilter result = new EFFilter(); //参数值集合 List<object> listArgs = new List<object>(); string filter = "1=1"; #region "处理动态过滤条件" if (gridParam.FilterList != null && gridParam.FilterList.Count > 0) { StringBuilder sb = new StringBuilder(); int paramCount = 0; DateTime dateTime; //操作符 string strOperator = string.Empty; foreach (var item in gridParam.FilterList) { //字段名称为空则跳过 if (string.IsNullOrEmpty(item.FieldName)) { continue; } //匹配枚举，防止SQL注入 Operator operatorEnum = (Operator)Enum.Parse(typeof(Operator), item.Operator, true); //跳过字段值为空的 if (operatorEnum != Operator.Null && operatorEnum != Operator.NotNull && string.IsNullOrEmpty(item.FieldValue)) { continue; } strOperator = operatorEnum.GetDescription(); if (item.IgnoreCase && !item.IsDateTime) { //2016-07-19添加查询时忽略大小写比较 item.FieldValue = item.FieldValue.ToLower(); item.FieldName = string.Format("{0}.ToLower()", item.FieldName); } switch (operatorEnum) { //等于,不等于，小于，大于，小于等于，大于等于 case Operator.EQ: case Operator.NE: case Operator.GT: case Operator.GE: case Operator.LT: case Operator.LE: if (item.IsDateTime) { if (DateTime.TryParse(item.FieldValue, out dateTime)) { if (!item.FieldValue.Contains("00:00:00") && dateTime.ToString("HH:mm:ss") == "00:00:00") { if (operatorEnum == Operator.LE) { listArgs.Add(DateTime.Parse(dateTime.ToString("yyyy-MM-dd") + " 23:59:59")); } else { listArgs.Add(dateTime); } } else { listArgs.Add(dateTime); } sb.AppendFormat(" AND {0} {1} @{2}", item.FieldName, strOperator, paramCount); } } else { listArgs.Add(ConvertToType(item.FieldValue, GetPropType<T>(item.FieldName))); sb.AppendFormat(" AND {0} {1} @{2}", item.FieldName, strOperator, paramCount); } paramCount++; break; case Operator.Like: case Operator.NotLike: case Operator.LLike: case Operator.RLike: listArgs.Add(item.FieldValue); if (operatorEnum == Operator.Like) { sb.AppendFormat(" AND {0}.Contains(@{1})", item.FieldName, paramCount); } else if (operatorEnum == Operator.NotLike) { sb.AppendFormat(" AND !{0}.Contains(@{1})", item.FieldName, paramCount); } else if (operatorEnum == Operator.LLike) { sb.AppendFormat(" AND {0}.EndsWith(@{1})", item.FieldName, paramCount); } else if (operatorEnum == Operator.RLike) { sb.AppendFormat(" AND {0}.StartsWith(@{1})", item.FieldName, paramCount); } paramCount++; break; case Operator.Null: listArgs.Add(item.FieldValue); sb.AppendFormat(" AND {0}=null", item.FieldName); paramCount++; break; case Operator.NotNull: listArgs.Add(item.FieldValue); sb.AppendFormat(" AND {0}!=null", item.FieldName); paramCount++; break; case Operator.In: sb.AppendFormat(" AND ("); foreach (var schar in item.FieldValue.Split(',')) { listArgs.Add(schar); sb.AppendFormat("{0}=@{1} or ", item.FieldName, paramCount); paramCount++; } sb.Remove(sb.Length - 3, 3); sb.AppendFormat(" )"); break; case Operator.NotIn: sb.AppendFormat(" AND ("); foreach (var schar in item.FieldValue.Split(',')) { listArgs.Add(schar); sb.AppendFormat("{0}!=@{1} and ", item.FieldName, paramCount); paramCount++; } sb.Remove(sb.Length - 3, 3); sb.AppendFormat(" )"); break; } if (sb.ToString().Length > 0) { filter = sb.ToString().Substring(4, sb.Length - 4); } } #endregion } result.Filter = filter; result.ListArgs = listArgs; return result; }
DataTable行转列
该部分是根据网友反馈后期再补充上的内容，意在完善行转列。下面给出实现代码
using Newtonsoft.Json; using System; using System.Collections.Generic; using System.Data; using System.Linq; namespace DataTable_RowToColumn { class Program { static void Main(string[] args) { DataTable dt = InitTable(); List<string> DimensionList = new List<string>() { "Area", "Month" }; string DynamicColumn = "Name"; List<string> AllDynamicColumn = null; DataTable dtResult = RowToColumn(dt, DimensionList, DynamicColumn, out AllDynamicColumn); Console.WriteLine(JsonConvert.SerializeObject(dtResult, Formatting.Indented)); Console.Read(); } /// <summary> /// 动态Linq方式实现行转列 /// </summary> /// <param name="list">数据</param> /// <param name="DimensionList">维度列</param> /// <param name="DynamicColumn">动态列</param> /// <returns>行转列后数据</returns> private static DataTable RowToColumn(DataTable dt, List<string> DimensionList, string DynamicColumn, out List<string> AllDynamicColumn) { //获取所有动态列 AllDynamicColumn = new List<string>(); foreach (DataRow dr in dt.DefaultView.ToTable(true, DynamicColumn).Rows) { if (dr[DynamicColumn] != null && !string.IsNullOrEmpty(dr[DynamicColumn].ToString())) { AllDynamicColumn.Add(dr[DynamicColumn].ToString()); } } //数值列 Dictionary<string, Type> AllNumberColumn = new Dictionary<string, Type>(); foreach (DataColumn item in dt.Columns) { if (item.DataType == typeof(int) || item.DataType == typeof(double) || item.DataType == typeof(float)) { AllNumberColumn.Add(item.ColumnName, item.DataType); } } //结果DataTable创建 DataTable dtResult = new DataTable(); foreach (var item in DimensionList) { dtResult.Columns.Add(item, typeof(string)); } //动态列 foreach (var dynamicValue in AllDynamicColumn) { foreach (var item in AllNumberColumn.Keys) { dtResult.Columns.Add(item + dynamicValue, AllNumberColumn[item]); } } //分组-优化性能 Dictionary<string, List<DataRow>> dict = new Dictionary<string, List<DataRow>>(); List<DataRow> drList = null; string groupKey = ""; foreach (DataRow dr in dt.Rows) { groupKey = ""; foreach (var item in DimensionList) { groupKey += dr[item] + "#"; } if (!dict.TryGetValue(groupKey, out drList)) { drList = new List<DataRow>(); dict[groupKey] = drList; } drList.Add(dr); } DataRow drReult = null; DataTable dtTemp = null; Dictionary<object, DataTable> dictTable = null; foreach (var kv in dict) { drReult = dtResult.NewRow(); var arrKey = kv.Key.Split('#'); int i = 0; foreach (var key in DimensionList) { drReult[key] = arrKey[i]; i++; } dictTable = (from p in kv.Value.AsEnumerable() group p by p.Field<object>(DynamicColumn) into g select g).ToDictionary(e => e.Key, e => e.CopyToDataTable()); foreach (var dynamicValue in AllDynamicColumn) { if (dictTable.TryGetValue(dynamicValue, out dtTemp)) { foreach (var numColumn in AllNumberColumn.Keys) { drReult[numColumn + dynamicValue] = dtTemp.Compute("sum(" + numColumn + ")", ""); } } else { foreach (var numColumn in AllNumberColumn.Keys) { drReult[numColumn + dynamicValue] = 0; } } } dtResult.Rows.Add(drReult); } return dtResult; } private static DataTable InitTable() { DataTable dt = new DataTable(); dt.Columns.Add("Name", typeof(string)); dt.Columns.Add("Area", typeof(string)); dt.Columns.Add("Month", typeof(string)); dt.Columns.Add("DfMoney", typeof(double)); dt.Columns.Add("SfMoney", typeof(double)); dt.Columns.Add("RqfMoney", typeof(double)); DataRow row = dt.NewRow(); row["Name"] = "张三"; row["Month"] = "2016-01"; row["Area"] = "江夏区"; row["DfMoney"] = 240.9; row["SfMoney"] = 30; row["RqfMoney"] = 25; dt.Rows.Add(row); row = dt.NewRow(); row["Name"] = "张三"; row["Month"] = "2016-02"; row["Area"] = "江夏区"; row["DfMoney"] = 167; row["SfMoney"] = 24.5; row["RqfMoney"] = 17.9; dt.Rows.Add(row); row = dt.NewRow(); row["Name"] = "小燕子"; row["Month"] = "2016-01"; row["Area"] = "江夏区"; row["DfMoney"] = 340.9; row["SfMoney"] = 20; row["RqfMoney"] = 55; dt.Rows.Add(row); row = dt.NewRow(); row["Name"] = "小燕子"; row["Month"] = "2016-02"; row["Area"] = "江夏区"; row["DfMoney"] = 67; row["SfMoney"] = 64.5; row["RqfMoney"] = 77.9; dt.Rows.Add(row); row = dt.NewRow(); row["Name"] = "李四"; row["Month"] = "2016-01"; row["Area"] = "洪山区"; row["DfMoney"] = 56.7; row["SfMoney"] = 24.7; row["RqfMoney"] = 13.2; dt.Rows.Add(row); row = dt.NewRow(); row["Name"] = "李四"; row["Month"] = "2016-02"; row["Area"] = "洪山区"; row["DfMoney"] = 65.2; row["SfMoney"] = 18.9; row["RqfMoney"] = 14.9; dt.Rows.Add(row); row = dt.NewRow(); row["Name"] = "尔康"; row["Month"] = "2016-01"; row["Area"] = "洪山区"; row["DfMoney"] = 156.7; row["SfMoney"] = 124.7; row["RqfMoney"] = 33.2; dt.Rows.Add(row); row = dt.NewRow(); row["Name"] = "尔康"; row["Month"] = "2016-02"; row["Area"] = "洪山区"; row["DfMoney"] = 35.2; row["SfMoney"] = 28.9; row["RqfMoney"] = 44.9; dt.Rows.Add(row); return dt; } } }
总结
本篇通过行转列引出了System.Linq.Dynamic，并且介绍了过滤功能，其实它的用处还有很多，等待大家发掘。下面给出本文示例代码：DynamicLinq
我的博客即将搬运同步至腾讯云+社区，邀请大家一同入驻：https://cloud.tencent.com/developer/support-plan