CSV Helper
Read, manipulate and write CSV files from urls or ingested via email
Last updated
Read, manipulate and write CSV files from urls or ingested via email
Last updated
With the CSV Helper, one can read, manipulate and write CSV files from URLs or other sources.
Authentication on the CSV-Helper was deprecated in 2021. Please use a REST Helper instead to fetch a file and then read it via CSV Helper.
This action reads a CSV string from a remote FTP location or HTTP URL location, including authorization tokens in headers.
Read produces a list (one entry for each row) of lists (one entry for each column), so the result of read_csv
might look like this:
This action creates a CSV string from previously generated data and saves it as a .csv
file. The user input contains a data reference—a ref ID pointing to a list of values—and a list of column field names. These column field names are used when iterating through the data given by the ref ID to extract values from the list element. If the columns are empty, then the CSV will contain just one column, namely, the corresponding line's string representation.
The input needs to be either a list of dictionaries [{...}, {....}]
or list of lists [[...], [...]]
If you only have one dictionary that you would like to insert, wrap it inside a list with the Dict Helper, e.g. [ {{ my_dictionary }} ]
This action produces a CSV string and uploads it to the designated S3 bucket for storage. It can also be written to a file on an FTP server or sent as a mail attachment.
Note: If you provide a list of dictionaries where some dictionaries contain empty keys (""
), the resulting CSV may contain null
values for those fields, as there are no valid column headers to map the data. To avoid this, ensure all dictionaries have valid, non-empty keys when generating the CSV file.
Example:
List of Lists:
Here, the last column has an empty header, so it is assigned a null
value in the resulting CSV.
List of Dictionaries:
In this case, the dictionaries have an empty key (""
), leading to empty values in the CSV output for that column.
Configuration outcome matrix:
Print_headers
selected_columns
new_column_names
outcome
yes
Some valid columns
Some valid names
If selected and new names have different length -> error Use new column names
yes
Some valid columns
nothing
Use selected column names
yes
nothing
Some valid names
Expect new_colum_names to have ONE entry
yes
nothing
nothing
Error
no
Some valid columns
Some valid names
Info(Unused new names)
no
Some valid columns
nothing
Regular csv without headers
no
nothing
Some valid names
Info(Unused new names)
no
nothing
nothing
CSV with one column and no headers
This action takes a string and converts it to CSV.
For instance, you want to transform "John,john@example.com,555-1234"
to CSV.
For all CSV actions, an encoding can be selected which should be used to write/read the CSV file.
This can be left at utf_8
for almost all cases. This is only needed if the CSV file was saved or is expected to be saved in a specific encoding.
This is especially relevant for umlaute and special characters.
If an encoding does not support characters used in the CSV (e.g. languages with incomplete coverage in latin-1), the step will result in an error.
The full list of supported encodings can be seen here.
To filter a CSV file, use Filter List V2 from the Dict Helper
To add a filter:
Connect to a Dict Helper step
Select the Filter List V2 action.
Fill the List reference input with a CSV reference without the {{ }}
Set up your filter criteria
Oftentimes, formats of e.g. dates are different in every system and need to be adjusted and modified to reflect the format of a different system.
Imagine a CSV with the column header datetime and the below three datetimestamps. That needs to be changed to only dates, like 2019-06-18
To do the above, you need to use a combination of:
The connector or helper passing the object
A looper that changes something in every object
and a dict_helper that changes a particular field in a dict
and a date helper that formats data in a particular way