Lectures

We break down practical CSV “gotchas” in Go: why naive parsing via strings.Split breaks on commas and newlines inside quotes, how quotes and escaping work ( ""), why a “file line” is not always the same as a CSV record, how to live with delimiters ,/ ;, why you need a header, and how a BOM ( "\ufeff") can silently break the first column name.

We break down why CSV cannot be reliably parsed with strings.Split, and how to work with the format correctly via the encoding/csv package. We learn to read records with csv.NewReader (modes Read() and ReadAll(), handling io.EOF, settings Comma, TrimLeadingSpace, FieldsPerRecord) and to write with csv.NewWriter (required Flush() and checking Error(), settings Comma and UseCRLF). At the end, we assemble a mini import/export of tasks via io.Reader/ io.Writer and lock in the typical mistakes.

We break down why successfully reading a CSV via csv.Reader still does not mean the data is valid, how to separate format errors from contract errors, and how to build a human-friendly report: record number ( rowNo), column, value, and reason. We will implement the helper getField, the error type RowError, validations requiredString/ parseInt/ parseBool, and importing from io.Reader while accumulating per-row issues.

We break down how to make CSV import “friendly”: instead of return on the first issue, we collect errors per row and return one final error via errors.Join. We learn to add context (row/column/value) via our own RowError type with Unwrap, carefully combine []error, unwrap a joined error back into a list, and check causes via errors.Is/ errors.As.

Let’s break down how to do reliable CSV export and import in Go: why you must not write directly to "tasks.csv", why you need the temp → rename protocol, how to finish writing correctly via Flush() + checking w. Error(), why it’s important to check the Close() error, and how to add extra insurance with a ".bak" backup. At the end: common mistakes and “checkpoints” where failures actually surface.