When a file can't be read

Some files can't be converted as they are, and some values don't fit the type their column gets. IntoSQL says what failed, where, and how to fix it. The same codes appear in the converter, the command line, the desktop app, the API (in a conversion's report) and MCP.

A problem that stops the whole file, such as a damaged workbook, ends the conversion. A problem with one value only affects its row: the row is left out and listed among the rejected rows, and the rest converts.

Reading the file#

unsupported_format#

"notas.pdf" isn't a spreadsheet, CSV or JSON file IntoSQL can read.

What to do: Save it as .xlsx or .csv and try again.

encrypted_file#

"nomina.xlsx" is protected with a password.

What to do: Open it, remove the password, save it and try again.

corrupt_file#

"ventas.xlsx" is damaged and can't be read to the end.

What to do: Open it in the app that made it and save a fresh copy.

empty_file#

"ventas.csv" has no rows.

What to do: Check that you picked the right file or sheet.

sheet_not_found#

There's no sheet called "Q4".

What to do: Pick one of: Q1, Q2, Q3.

other_table_skipped#

Row 58 starts another table, which was left out.

What to do: Convert it on its own: copy that range to a new sheet or file.

invalid_json_row#

Row 7 isn't valid JSON, so it was left out.

What to do: Check that each line holds one complete JSON object.

ragged_row#

Row 212 has 7 values but the header has 6.

What to do: Missing values are left empty and extra values are dropped. Check the row if that isn't what you want.

Values#

invalid_date#

Row 418, column "Alta": 31/02/2024 isn't a date.

What to do: Keep the value as text, or leave the row out.

invalid_number#

Row 77, column "Saldo": pendiente isn't a number.

What to do: Keep the column as text, or correct the value.

invalid_time#

Row 15, column "Hora": 25:10 isn't a time of day.

What to do: Keep the column as text, or correct the value.

invalid_boolean#

Row 30, column "Activo": tal vez isn't yes or no.

What to do: Keep the column as text, or correct the value.

value_too_long#

Row 9, column "Notas" is longer than SQL Server allows (4000 characters).

What to do: Use a text type without a limit, or shorten the value.

ambiguous_reading#

Check "Fecha": 03/02/2024 could mean February 3 or March 2.

What to do: If February 3 is wrong, change how the column is read.

mixed_number_formats#

"Importe" mixes 1.234,50 and 1,234.50, so it stays text.

What to do: Pick a number type for the column to read every value as a number.

Options#

table_name_empty#

The table name is empty.

What to do: Type a name, or leave the field out to use the file's name.

table_name_too_long#

"ventas_mensuales_por_sucursal_y_region" is longer than Oracle allows (30 characters).

What to do: Shorten the name to 30 characters or fewer.

column_name_empty#

The new name for column 2 is empty.

What to do: Type a name, or leave the column's name as detected.

column_name_too_long#

"fecha_de_la_ultima_actualizacion_del_registro" is longer than Oracle allows (30 characters).

What to do: Shorten the name to 30 characters or fewer.

column_not_found#

There's no column 9 to change: the last column is 6.

What to do: Check the column numbers: they start at 1.

no_columns#

Every column is left out, so there's nothing to write.

What to do: Keep at least one column.

batch_rows_out_of_range#

Rows per batch must be between 1 and 10000; it is 0.

What to do: Pick a number in that range. 500 suits most databases.

sample_rows_too_small#

Type inference needs at least 100 sample rows; it is 10.

What to do: Raise the sample size to 100 or more.

unsupported_version#

This request uses contract version 2; this engine reads version 1.

What to do: Update IntoSQL, or send a request made for version 1.