用于解析 CSV 文本行的 Linq Lookup

Linq Lookup to parse a CSV text line

问题

前一段时间我曾问过 this question,从那时起要求发生了一些变化。

现在,可能会有一个包含以下行的文件:

Bryar22053;ADDPWN;Bryar.Suarez@company.com;ACTIVE
Nicole49927;ADDPWN;Nicole.Acosta@company.com;ACTIVE
Rashad58323;ADDPWN;Rashad.Everett@company.com;ACTIVE

取第一行。跳过第一个值 Bryar22053 并使用相同的查找:

var columnCount = dataRow.Skip(1).Count();
var modular = 0;

// Simple Enum
var rightsFileType = new RightsFileType();

if (columnCount % 2 == 0)
{
    rightsFileType = RightsFileType.WithoutStatus;
    modular = 2;
}
else if (columnCount % 3 == 0)
{
    rightsFileType = RightsFileType.WithStatus;
    modular = 3;
}

var lookup = dataRow.Skip(1).Select((data, index) => new
{
    lookup = index % modular,
    index,
    data
}).ToLookup(d => d.lookup);

查找对象现在有三组:

> ? lookup[0].ToList() Count = 1
>     [0]: { lookup = 0, index = 0, data = "ADDPWN" } ? lookup[1].ToList() Count = 1
>     [0]: { lookup = 1, index = 1, data = "Bryar.Suarez@company.com" } ? lookup[2].ToList() Count = 1
>     [0]: { lookup = 2, index = 2, data = "ACTIVE" }

如果是原来的情况,它只是 System1、User1、System2、User2... lookup 将有两个组,下面的代码将起作用:

List<RightObjectRetrieved> rights;
rights = lookup[0].Join(lookup[1], system => system.index + 1, username => username.index, (system, username) => new
{
    system = system.data,
    useraname = username.data
}).Where(d => !string.IsNullOrEmpty(d.system)).Select(d => new RightObjectRetrieved {UserIdentifier = userIdentifier, SystemIdentifer = d.system, Username = d.useraname, RightType = rightsFileType}).ToList();

// rights => Key = System Identifier, Value = Username  

但是第三个 'status' 为 System1,User1,Status1,System2,User2,Status2...,我在尝试加入并获取全部三个时遇到问题。请帮忙。

编辑 这是我的原始数据:

// Method has parameter localReadLine (string) that has this:
// Bryar22053;ADDPWN;Bryar.Suarez@company.com;ACTIVE

// Data line
var dataRow = localReadLine.Split(new[] { ToolSettings.RightsSeperator }, StringSplitOptions.None);

// Trim each element
Array.ForEach(dataRow, x => dataRow[Array.IndexOf(dataRow, x)] = x.Trim());

到目前为止已尝试(失败)

rights = lookup[0].Join(lookup[1], system => system.index + 1, username => username.index, status => status.index, (system, username, status) => new
{
    system = system.data,
    useraname = username.data,
    status = status.data
}).Where(d => !string.IsNullOrEmpty(d.system)).Select(d => new RightObjectRetrieved {UserIdentifier = userIdentifier, SystemIdentifer = d.system, Username = d.useraname, RightType = rightsFileType}).ToList();

rights = lookup[0].Join(lookup[1], system => system.index + 1, username => username.index, (system, username) => new
{
    system = system.data,
    useraname = username.data
}).Join(lookup[2], status => status.index, (status) => new
{
    status = status.data
}).Where(d => !string.IsNullOrEmpty(d.system)).Select(d => new RightObjectRetrieved {UserIdentifier = userIdentifier, SystemIdentifer = d.system, Username = d.useraname, RightType = rightsFileType, Status = ParseStatus(status)}).ToList();

我认为您需要稍微拆分一下您的实施。

让我们声明一个 class 来保存数据:

class Data
{
    public string System { get; set; }
    public string Username { get; set; }
    public string Status { get; set; }
}

现在,让我们定义几个解析函数来解析一行。 第一个将解析包含状态的行:

var withStatus = (IEnumerable<string> line) => line
    .Select((token, index) => new { Value = token, Index = index })        
    .Aggregate(
        new List<Data>(),
        (list, token) =>
        {
            if( token.Index % 3 == 0 )
            {
                list.Add(new Data { System = token.Value });
                return list;
            }
            var data = list.Last();
            if( token.Index % 3 == 1 )
                data.Username = token.Value;
            else
                data.Status = token.Value;
            return list;
        });

第二个将解析不包含状态的行:

var withoutStatus = (IEnumerable<string> line) => line
    .Select((token, index) => new { Value = token, Index = index })
    .Aggregate(new List<Data>(),
        (list, token) => 
        {
            if( token.Index % 2 == 0)
                list.Add(new Data { System = token.Value });
            else
                list.Last().Username = token.Value;
            return list;
        });

准备好所有这些后,您将需要以下内容:

  1. 确定模数
  2. 迭代文件的行并解析每一行
  3. 分组并汇总结果

剩余的代码如下所示:

var lines = streamReader.ReadAllLines(); // mind the system resources here!
var parser = lines.First().Split(';').Length % 2 == 0 ? withoutStatus : withStatus;
var data = lines.Skip(1) // skip the header
    .Select(line =>
    {
        var parts = line.Split(';');
        return new
        {
            UserId = parts.First(),
            Data = parser(parts.Skip(1))
        };
    })
    .GroupBy(x => x.UserId)
    .ToDictionary(g => g.Key, g => g.SelectMany(x => x.Data));

现在您有一个 Dictionary<string, Data>,其中包含用户 ID 及其信息。

当然,更优雅的解决方案是将每个解析函数分离到它自己的 class 中,然后将这些 class 加入一个通用接口下,以防需要添加更多信息未来,但上面的代码应该可以工作,并让您了解应该做什么。

如果您想使用联接:

var result = lookup[0]
    .Join(lookup[1],
        system => system.index,
        username => username.index - 1,
        (system, username) => new {system = system.data, username = username.data, system.index})
    .Join(lookup[2],
        d => d.index,
        status => status.index - 2,
        (d, status) => new {d.system, d.username, status = status.data})
    .ToList();

按记录分组的另一种选择,仅 select 来自它的数据(从我的角度来看看起来更具可读性):

var result = dataRow
    .Skip(1)
    .Select((data, index) => new {data, record = index / 3})
    .GroupBy(r => r.record)
    .Select(r =>
    {
        var tokens = r.ToArray();
        return new
        {
            system = tokens[0].data,
            username = tokens[1].data,
            status = tokens[2].data
        };
    })
    .ToList();