XML to CSV / Excel Converter

About the XML to CSV / Excel Converter

Get the records out of an XML file and into a spreadsheet. Open or paste XML — an export, a product or job feed, RSS or Atom, a sitemap, KML places, a SOAP response — and the element that repeats is found for you (/catalog/book, /rss/channel/item, /urlset/url): each one becomes a row, and everything inside it becomes columns named by their path — attributes as id or price/@currency, child elements as title or author/name. Children that repeat inside a record (several authors, several categories) are joined in one cell or get a column each. With a pass the table saves as CSV, TSV or Excel; the free preview shows its first rows.

It works the other way too: a CSV or Excel table becomes XML, with the root and row element names you choose, columns as child elements or attributes (@id columns are always attributes), nested elements from column names such as address/city, and every value escaped as XML requires. Column names that are not valid XML names are made valid, and the page says which.

Files of any size are read piece by piece in your browser (gzip-compressed ones too), and mistakes in the XML are listed with their line and column. Without a pass, the free preview shows the structure and the first rows of the table, or the first lines of the XML; the files and copying unlock with a Pro pass.

Common uses

• Opening a product feed, an export or an API response in Excel.
• Turning RSS or Atom feeds, sitemaps or KML placemarks into a list.
• Comparing or filtering XML data in a spreadsheet, then writing it back as XML.
• Making the XML file an importer asks for from a spreadsheet.

How the rows and columns are found

The file is read once to count every element path: how often it occurs, how often it repeats under one parent, and which children and attributes it has. The element that repeats and holds the most fields becomes the row — usually what a person would pick — and the others are offered in the list. Paths use XPath’s short notation: /catalog/book starts at the root, //item means an item at any depth (XPath 1.0).

Inside each record, every attribute and every element with text is a column named by its path from the record. The record element’s own attributes have plain names (id) unless a child element has the same name (then @id). An element that has both text and child elements keeps its own text in a #text column. Spaces around text are trimmed, as XML files are usually indented.

What is checked in the XML

The reader follows XML 1.0: one root element, matching start and end tags, attribute values in quotes and each attribute once per tag, no < inside a value, the five predefined entities (< > & ' "), character references such as  , and entities the document’s DOCTYPE declares. Mistakes are listed with their line and column, and the rest of the file is still read where it can be. Nothing is ever fetched: external DTDs and entities are not loaded, and documents whose entities expand without limit are refused.

Writing XML from a table

Each row becomes one element under the root. Columns are child elements or attributes; a column whose name starts with @ is an attribute, and address/city writes a <city> inside an <address>. Values are escaped (& < > in text, also " and line breaks in attributes), characters XML 1.0 does not allow are left out, and names that are not valid XML names — 1st name, my field — become valid ones (_1st_name, my_field). The CSV side follows RFC 4180.

Limitations

• XML files of any size are read; each record is read whole, so a single record must fit in memory. Gzip files are unpacked as they are read.
• Records are flat rows: a deeply nested record gives many columns, and lists inside lists are joined. Mixed text such as Hello <b>world</b> keeps its own text and its children’s text in separate columns.
• Only internal DOCTYPE entities are read; DTD validation, XML Schema, XInclude and namespaces beyond their prefixes are not applied.
• Tables turned into XML are read whole: CSV files up to 100 MB and workbooks up to 50 MB.
• Excel sheets hold at most 1,048,576 rows and 16,384 columns; choose CSV for more.

Privacy

Everything happens in your browser. What you enter or open here is not uploaded or stored by MySmartCoPilot.