Encoding: pick data encoding for your file. Encoding determines how characters are stored in text files and picking incorrect encoding can result in incorrectly read data (for example, some accented letters may show incorrectly). Most common encodings are UTF-8 (universal encoding that can handle any characters), windows-1252 (US English encoding commonly used for files created on Windows) or ISO-8859-1 (US ASCII encoding).

Delimiter: allows you to select column delimiter. Column delimiter is a single character that serves as a separator between columns in the file. Most common delimiters are comma (",") and tab. You can also type your own delimiter if none of the default delimiters work for you. Setting correct delimiter ensures that all columns are read from the file as expected.

Quoting: determines whether values in your source file are enclosed in quotes or not. You can select from three options - double quotes, single quotes or N/A (no quotes).

First line is a header: tell the CSV connector that it should interpret first line of the data as header and read names of columns from it. If this is unchecked, columns in your data set will have sequentially generated names like field1, field2, and so on.

Files of any size. There is no limit of the total number of rows in your file. Note that by default only the first 1000 rows are shown in the preview. To ensure the whole file is processed, run the job once you are finished building the transformation.

Files with any number of columns. While there is no technical limitation to how many columns can you files have, Wrangler implements a soft limit set to 1000 columns by default to help you prevent cases where your browser may not be able to handle display of too large of a sample. You can change this limit by contacting your IT and asking them to increase value of Wrangler.MAX_NUMBER_OF_COLUMNS property.

Files with varying number of columns on different rows. You files do not have to have the same number of columns on each rows. However, this feature works correctly only when the header or one of the sample rows is the longest row in the file. For example, the connector will work fine if your header contains 20 columns but some of the rows in your file contain fewer than 20 columns. By default, connector will read the first 512 kB of your data or first 1000 rows whichever comes first to determine how many columns does your data have. You can work with your IT team to increase these limits by setting CSVAnalyzer.BYTES_TO_ANALYZE and CSVAnalyzer.LINES_TO_ANALYZE settings (see more here).

Files in which the longest row (a row with the largest number of columns) does not appear within the analyzed sample. For example, if you have a file that has a header with 20 columns and much later in the file (after first 1000 rows) you have a record with more than 20 columns. See above for more details about increasing sampling limits.

It does not support files with multiple different types of column delimiters. If you have such a file, you will need to pick one of the delimiters and then split the columns using the other delimiter in your Wrangler job using Calculate formula step.