Question

因此，我试图制作一个备份脚本，该脚本将从我的mssql下载CSV，然后压缩文件，然后将备份上传到Amazon S3。

我每天运行脚本时，表的问题是平均有2000万行。看起来就像永远滞后，直到20分钟后完成。我想知道是否有办法专门显示invoke-sqlcmd的进度条。我做了一些研究，我能找到的所有示例都是仅在for循环上制作进度条，而不是对单个命令进行进度。

这是我的代码：

ECHO "Starting Download"
    Import-Module sqlps
    #$SQLquery="SELECT * FROM dbo.$PREFIX$i"
    $SQLquery="SELECT * FROM dbo.events"
    ECHO "Executing query = $SQLquery"

    $hostname = "."
    $pass = "test"
    $usern = "test"
    $database = "theDB"

    $result=invoke-sqlcmd -ServerInstance $hostname  -query $SQLquery -HostName $hostname -Password $pass -Username $usern -Database $database -verbose


    #echo $result
    pause
    $result |export-csv -path $CSVPATH -notypeinformation
    pause
    ECHO "Starting Zip:"
    Compress-Archive -LiteralPath $CSVPATH -CompressionLevel Optimal -DestinationPath $ZIPPATH

    ECHO "Starting Delete: $CSVPATH "
    del "$CSVPATH" 
    echo "Removed $CSVNAME" 

    aws s3 cp $ZIPPATH s3://test_$ZIPNAME

    pause

此脚本有效，但正如我所说，我想在invoke-sqlcmd中添加一个进度条，以使其在下载巨大文件时不会看起来像冻结的一样。

这是我到目前为止可以找到的，但这仅适用于循环进度

$VerbosePreference = "Continue"
Write-Verbose "Test Message"

for ($a=1; $a -lt 100; $a++) {
  Write-Progress -Activity "Working..." -PercentComplete $a -CurrentOperation "$a% complete" -Status "Please wait."
  Start-Sleep -Milliseconds 100
}

Answer 1

考虑到约2000万个庞大的记录数据集，使用System.Data.Common namespace中的某些.NET类可能是一个好主意。而且我不确定Export-Csv的实现方式，但是System.IO.StreamWriter对于写入大文件非常有效。

带有内联注释的简单测试/工作示例：

# replace $tableName with yours
$sqlCount = "SELECT COUNT(*) FROM dbo.$($tableName)";
$sqlSelect = "SELECT * FROM dbo.$($tableName)";
$provider = [System.Data.Common.DbProviderFactories]::GetFactory('System.Data.SqlClient');
$connection = $provider.CreateConnection();
# replace $connectionString with yours, e.g.:
# "Data Source=$($INSTANCE-NAME);Initial Catalog=$($DATABASE-NAME);Integrated Security=True;"; 
$connection.ConnectionString = $connectionString;
$command = $connection.CreateCommand();

# get total record count for Write-Progress
$command.CommandText = $sqlCount;
$connection.Open();
$reader = $command.ExecuteReader();
$totalRecords = 0;
while ($reader.Read()) {
    $totalRecords = $reader[0];
}
$reader.Dispose();

# select CSV data
$command.CommandText = $sqlSelect;
$reader = $command.ExecuteReader();

# get CSV field names
$columnNames = @();
for ($i = 0; $i -lt $reader.FieldCount; $i++) {
    $columnNames += $reader.GetName($i);
}

# read and populate data one row at a time
$values = New-Object object[] $columnNames.Length;
$currentCount = 0;
# replace $CSVPATH with yours
$writer = New-Object System.IO.StreamWriter($CSVPATH);
$writer.WriteLine(($columnNames -join ','));
while ($reader.Read()) {
    $null = $reader.GetValues($values);
    $writer.WriteLine(($values -join ','));
    if (++$currentCount % 1000 -eq 0) {
        Write-Progress -Activity 'Reading data' `
            -Status "Finished reading $currentCount out of $totalRecords records." `
            -PercentComplete ($currentCount / $totalRecords * 100);
    }
}
$command.Dispose();
$reader.Dispose();
$connection.Dispose();
$writer.Dispose();

用于invoke-sqlcmd的Powershell进度栏

1 个答案: