125 lines
3.7 KiB
C#
125 lines
3.7 KiB
C#
using System.Text.RegularExpressions;
|
||
using ZstdSharp;
|
||
|
||
namespace MesETL.App.Helpers;
|
||
|
||
public static partial class DumpDataHelper
|
||
{
|
||
[GeneratedRegex(@"'.+\.dat'")]
|
||
private static partial Regex MatchDatFile();
|
||
[GeneratedRegex(@"\([^)]*\)")]
|
||
private static partial Regex MatchBrackets();
|
||
|
||
/// <summary>
|
||
/// 从MyDumper导出的SQL文件内容中读取表头
|
||
/// </summary>
|
||
/// <param name="content"></param>
|
||
/// <returns></returns>
|
||
/// <exception cref="ArgumentException"></exception>
|
||
public static string[] GetCsvHeadersFromSqlFile(string content)
|
||
{
|
||
var match = MatchBrackets().Match(content);
|
||
if (!match.Success)
|
||
throw new ArgumentException("输入的SQL内容有误，无法提取表头", nameof(content));
|
||
return ParseHeader(match.ValueSpan);
|
||
|
||
string[] ParseHeader(ReadOnlySpan<char> headerStr)
|
||
{
|
||
headerStr = headerStr[1..^1];
|
||
Span<Range> ranges = stackalloc Range[50];
|
||
var count = headerStr.Split(ranges, ',');
|
||
var arr = new string[count];
|
||
|
||
for (var i = 0; i < count; i++)
|
||
{
|
||
arr[i] = headerStr[ranges[i]].Trim("@`").ToString(); // 消除列名的反引号，如果是变量则消除@
|
||
}
|
||
|
||
return arr;
|
||
}
|
||
}
|
||
|
||
/// <summary>
|
||
/// 从MyDumper导出的Csv文件名解析出表名
|
||
/// </summary>
|
||
/// <param name="filePath"></param>
|
||
/// <returns></returns>
|
||
public static string GetTableNameFromCsvFileName(ReadOnlySpan<char> filePath)
|
||
{
|
||
filePath = filePath[(filePath.LastIndexOf('\\') + 1)..];
|
||
var firstDotIdx = -1;
|
||
var secondDotIdx = -1;
|
||
var times = 0;
|
||
for (var i = 0; i < filePath.Length; i++)
|
||
{
|
||
if (filePath[i] == '.')
|
||
{
|
||
++times;
|
||
if(times == 1)
|
||
firstDotIdx = i;
|
||
if (times == 2)
|
||
{
|
||
secondDotIdx = i;
|
||
break;
|
||
}
|
||
}
|
||
}
|
||
|
||
return filePath[(firstDotIdx+1)..secondDotIdx].ToString();
|
||
}
|
||
|
||
/// <summary>
|
||
/// 从MyDumper导出的SQL文件内容中读取CSV文件名
|
||
/// </summary>
|
||
/// <param name="txt"></param>
|
||
/// <param name="regex"></param>
|
||
/// <returns></returns>
|
||
public static Task<string[]> GetCsvFileNamesFromSqlFileAsync(string txt, Regex regex)
|
||
{
|
||
//var txt = await File.ReadAllTextAsync(filePath);
|
||
var matches = regex.Matches(txt);
|
||
return Task.FromResult(matches.Select(match => match.ValueSpan[1..^1].ToString()).ToArray());
|
||
}
|
||
|
||
/// <summary>
|
||
/// 检查字符串是否为16进制
|
||
/// </summary>
|
||
/// <param name="str"></param>
|
||
/// <returns></returns>
|
||
public static bool CheckHexField(string? str)
|
||
{
|
||
if (string.IsNullOrWhiteSpace(str))
|
||
return false;
|
||
|
||
if (str.StartsWith('\"'))
|
||
return false;
|
||
|
||
var isAllDigit = true;
|
||
|
||
foreach (var c in str)
|
||
{
|
||
if (!char.IsAsciiHexDigit(c))
|
||
return false;
|
||
if (!char.IsNumber(c))
|
||
isAllDigit = false;
|
||
}
|
||
|
||
if (isAllDigit) //避免全数字
|
||
return false;
|
||
|
||
return true;
|
||
}
|
||
|
||
/// <summary>
|
||
/// 将输入流以ZSTD标准解压为字符串
|
||
/// </summary>
|
||
/// <param name="stream"></param>
|
||
/// <returns></returns>
|
||
public static async Task<string> DecompressZstAsStringAsync(Stream stream)
|
||
{
|
||
await using var ds = new DecompressionStream(stream);
|
||
var reader = new StreamReader(ds);
|
||
return await reader.ReadToEndAsync();
|
||
}
|
||
|
||
} |