question

Yes you are correct. I just pushed a commit to my GitHub containing the replacing of 1 backslash to 2 backslashes when writing to CSV. This fixes the issue. I get exactly the same output as my input when i copy data in and export it with my tool.

Thanks!

If I add a backslash to all the backslashes in your csv I get the expected data. Are you seeing the debug view on Visual Studio? Visual Studio shows escaped strings in debug mode, you can click the magnifying glass icon to show the plain unescaped string.

Thats what I was thinking as well. However, when I change the single backslash to double backslashes as you suggested, i'm actually seeing two backslashes in my database. So I either get 0 or 2 backslashes. The real question is: how do I get just 1.

Also the suggestion you are making with using another escape character is going to be a challenge, since we can't predict which characters are used in our data. So the same issue will probably occur with another character then.

Looking forward to your thoughts.

Hi,

Thanks for your help so far. With your code i'm able to get satisfying results. However i'm working with existing tables. This is how to reproduce:

First of all create this keyspace & table:

CREATE KEYSPACE stresscql WITH replication = {'class': 'NetworkTopologyStrategy', 'DC1': '3', 'DC2': '3'}  AND durable_writes = true;

CREATE TABLE stresscql.insanitytest (
    mailboxid text,
    alias text,
    PRIMARY KEY (mailboxid, alias)
) WITH COMPACT STORAGE
    AND CLUSTERING ORDER BY (alias ASC)
    AND bloom_filter_fp_chance = 0.1
    AND caching = {'keys': 'ALL', 'rows_per_partition': 'NONE'}
    AND comment = 'A table of many types to test wide rows'
    AND compaction = {'class': 'org.apache.cassandra.db.compaction.LeveledCompactionStrategy'}
    AND compression = {'chunk_length_in_kb': '64', 'class': 'org.apache.cassandra.io.compress.LZ4Compressor'}
    AND crc_check_chance = 1.0
    AND dclocal_read_repair_chance = 0.1
    AND default_time_to_live = 0
    AND gc_grace_seconds = 864000
    AND max_index_interval = 2048
    AND memtable_flush_period_in_ms = 0
    AND min_index_interval = 128
    AND read_repair_chance = 0.0
    AND speculative_retry = '99PERCENTILE';

Next login to cqlsh and use this command:

COPY stresscql.insanitytest FROM '/myfolder/testdata.csv';

testdata.csv.zip

Note: when i change the single backslashes in the testdata.csv to double backslashes, I get double backslashes in my code while debugging. But when I have single backslashes (like the sample attached) I get zero backslashes in my code while debugging.

Next up: use my tool (which has the code i provided earlier) from https://github.com/RobHofmann/Cassandra.BackupAndRestore/. Make sure to edit appsettings.json with your values.

PS. I use a powershell script to spin up a 6 node cluster in Docker. You can find that script & info here: https://github.com/RobHofmann/Cassandra.LocalDockerCluster

Looking forward to your input!

I'm still not able to reproduce the issue with the code sample that you provided. Which driver and server versions are you using? I tested this with DataStax C# Driver for Apache Cassandra 3.11.0 and Apache Cassandra 3.11.4.

I inserted this row into the database ('cn' is of type 'text'):

INSERT INTO users (isdeleted,id,firstname,cn,lastname) VALUES (false,be8a976f-0616-405b-8d79-8d85915c4889,'3213','t*\x10r>\nw@S=','55');

And this is the resulting .csv file:

False,be8a976f-0616-405b-8d79-8d85915c4889,"3213","t*\x10r>\nw@S=","55"

This is the method i use for fetching rows:

public Task<RowSet> GetRowsAsync(ISession session, string table)
{
return session.ExecuteAsync(new SimpleStatement($"SELECT * FROM {table}"));
}

This is the main code:

_logger.Log("Fetching rows", fullTableName);
var dataRows = await _cassandraService.GetRowsAsync(session, fullTableName);
_logger.Log("Fetched rows", fullTableName);

_logger.Log("Start writing data to CSV", fullTableName);
TextWriter tw = new StreamWriter($"{_cassandraConfiguration.BackupTargetFolder}/{fullTableName}.csv");
_counter.Reset();
foreach (var dataRow in dataRows)
{
var stringRow = new string[dataRows.Columns.Length];
for (int i = 0; i < dataRows.Columns.Length; i++)
{
var data = dataRow.GetValue<object>(dataRows.Columns[i].Name);
if (data == null)
{
stringRow[i] = "";
continue;
}

if (data is DateTimeOffset)
stringRow[i] = ((DateTimeOffset)data).ToString("yyyy-MM-dd HH:mm:ss.fff+0000");
else if (data is string)
stringRow[i] = $"\"{data}\"";
else
stringRow[i] = data.ToString();
}
tw.WriteLine(string.Join(",", stringRow));
_counter.IncrementCounter();
}

tw.Flush();
tw.Close();

Let me know if something is unclear.

Hi, I'm not able to reproduce this, can you provide a code sample?