Common errors during upload

If you encounter issues uploading your data or data dictionary, check this page. If you cannot identify the error or do not know how to solve it, ask for help

Dataset Upload Errors

A few common errors are flagged during the data upload process. If you hit an error on the data preview page after selecting your datafile, check your dataset for the following errors:

• Your datafile is not a .csv file. For more information, see our FAQs

• Your datafile contains a column with an empty header (i.e. missing a value in the first row). You might have forgotten to type a name for a column or have an empty column in the middle of your dataset. Make sure every column of your dataset has a column name in the first row.

• Your datafile contains duplicate header names. Make sure every column must have a unique column header.

• One of your column headers has a variable name that is longer than 64 characters. There is a 64 character limit for the column headers; make sure every header is 64 characters or less.

• You used a comma in your datafile, and the .csv file is reading the comma as a separator between one entry and the next. This can be a source of error for header names and data entries. If this is the error, you will find rows that have different numbers of columns and/or entries that have been shifted to other columns.

  • You can fix this error by removing commas from the entries in your datafile.

  • A second way to fix this is to open your .csv in a spreadsheet software like Excel, correct for any shifted cells, and save as a new .csv through the program. Many software programs will naturally treat cells with commas as a fixed sequence of characters and won't treat those within-cell commas as separators between cells. This treatment will be maintained when you upload the new .csv to the ODC.

• Your datafile might be too large. If your dataset is larger than 100Mb or has a total number of cells larger than 3,000,000, see the section "What if my dataset is too large" below.

For more information about how to prepare your data for upload, see the getting your data ready.

Data Dictionary Upload Errors

When uploading a data dictionary file, there are a few possible errors that are flagged during the data dictionary upload process. You will be notified directly on the upload page; as a reference, the possible errors include:

• Your data dictionary is not a .csv file. The data dictionary must be a .csv file for upload.

• Your data dictionary is missing a column. The data dictionary must include all 9 required columns with exact spelling.

• Your data dictionary does not include an entry for every column of your dataset. Every Variable (i.e. column header) in your dataset must have a respective row in your data dictionary. Please ensure that the dataset variable names exactly match (spelling, caps) the variable names in the data dictionary.

• While not an error, if you have rows in your data dictionary that are missing values under Title or Description, this will flag a warning. This is not required for initial data dictionary upload, but during DOI request/dataset publication, we require that every row in your data dictionary have at least VariableName, Title, and Description columns filled out.

For more information on how to prepare your data dictionary for upload, see the data dictionary section.

What if my dataset is too large

If your dataset is too large (e.g., your dataset is larger than 100Mb or has a total number of cells larger than 3,000,000), it can cause an error during the data upload process. The error can also happen when you try to replace your dataset using the update a dataset function. In both cases, we recommend splitting up your dataset-to-be-uploaded into chunks with fewer rows and utilizing the append data option to add your dataset piece by piece.

Importantly, every chunk of your dataset must have the same column headers in the first row of each .csv file. Make sure you (1) split your dataset along the rows and not along the columns and (2) include the column headers in every file.