SureMatters

Guides · Doing the work

Getting a Microsoft Purview export into SureRedact

For any UK organisation preparing a subject access request from Microsoft 365 — usually the school business manager or DPO, with whoever runs eDiscovery for the organisation.

Last reviewed . Reviewed again when Microsoft changes Purview's export options.

Menus move; if a name below is not what you see, look for the nearest match rather than assuming the option has gone.


Before you export

Who does this. Exporting from Purview needs the eDiscovery role, which usually sits with IT or your Microsoft 365 provider rather than with the person handling the request. Microsoft’s documentation says the exporting user needs a Microsoft 365 Enterprise E3 or E5 licence; the education plans A3 and A5 are the equivalents — check with whoever administers your tenant.

Scope the search to the request. Search for the data subject, not for everything. The export you get is the export you will have to review, and a proportionate search is what the law now asks for — see Subject access requests in schools.

Keep the reports. Every Purview process produces a report package with Items.csv (every item, where it came from, when), Summary.csv and Settings.csv. Microsoft describes Settings.csv as “the authoritative record of the configuration used for an export”. Together they are your evidence of what you searched and how. Keep them with the request’s file. They are not part of what you send the requester.


The export settings

These are the options on Purview’s Export pane when you export from a search. The names are Microsoft’s. Where an option matters to SureRedact, the reason is given.

OptionChooseWhy
Select items to include in your exportIndexed items that match your search query and partially indexed items that might not match query”Partially indexed” means Microsoft could not read the item’s text — image-only PDFs, scans, photographs of letters. Those are exactly what a school’s records contain and exactly what SureRedact is for: it reads scanned pages with its own text reader, on every computer. Leaving them out leaves out the scans.
Organize conversation into HTML transcriptTickTeams chats arrive as readable HTML transcripts. SureRedact reads HTML; it does not read the individual chat messages Purview would otherwise export.
Access links (cloud attachments) in messagesTick, Latest version onlyA link to a file in OneDrive is an attachment in everything but form. Without this the message arrives and the file it pointed at does not.
Export typeExport items with items reportThe other choice exports the report only.
Export formatCreate .msg files for messagesNot Create PSTs for messages. SureRedact reads .msg files directly; it cannot open a PST. See “Email” below for the three things it will tell you.
Organize data from different locations into separate foldersTickOne folder per mailbox or site. You will want to know which is which.
Include folder and path of the sourceTickThe file’s original location travels with it — Inbox › Safeguarding › 2025. That is provenance: SureRedact records each document’s source path inside the Matter, for you. The disclosure pack carries no paths from your computer.
Condense paths to fit within 259 charactersTickDeeply nested folders produce paths Windows cannot unzip. Microsoft’s own advice; also the reason to avoid the built-in extractor.
Give each item a friendly nameTickFiles keep their real names and messages are named by their subject line. Without it every file is named by a long random identifier and you need Items.csv to know what anything is. One consequence: a subject line can itself name a person. The document’s name is what you and, in the pack, the requester will see — rename a file in the folder if its name needs care.

If you have Purview eDiscovery Premium and export from a review set instead, the same format and packaging choices appear, and the download window is 30 days rather than 14. Premium also offers to replace items you have redacted in Purview with its own redacted PDFs; if you are using SureRedact you will not be doing that.


Download and unzip

  1. When the export finishes, download every package listed under Export packages. Microsoft’s guidance is Edge, pop-ups allowed for the Purview site, and a local drive rather than a network share.
  2. Do it within 14 days. After that the export is deleted and cannot be recovered; you would have to run it again.
  3. Unzip with 7-Zip, WinZip or similar. Microsoft says to avoid the built-in Windows extractor, which fails on long paths and resets file dates. The original dates are inside the files and in Items.csv regardless.
  4. Move the Reports folder out of the export and into the request’s own file. It is your record, not the requester’s material.

Into SureRedact

Create a new Matter and point it at the unzipped export folder. SureRedact walks every subfolder, converts what it can to PDF and lists what it could not.

  • Free reads PDFs and email messages — the .msg files a Purview export contains, each with its attachments inside its PDF. Office documents, images, text and HTML files on their own are converted by Professional; on Free they are counted and left out, and you are told so.
  • What could not be included is always shown, as a count and a list, and the disclosure pack tells the requester the count. Nothing is silently dropped. For the file types below that it does not read, SureRedact says what to do beside the file.

What a Purview export contains, and what SureRedact does with it

In the exportSureRedactWhat to do
PDFs, scans, photographsReadsNothing
Word, Excel, PowerPoint — including the macro-enabled kindsConvertsNothing. Macros are never run; only the text is taken.
Email as .msg filesConverts — headers, body, attachments, and messages attached to messagesNothing. See “Email” below for the three notes it can leave.
Teams chats as HTML transcriptsConvertsNothing — if you ticked the transcript option
Text, CSV and Rich Text (.rtf) filesConvertsNothing
iPhone photos saved as .heicDoes not read yetSureRedact says so: save or export the image as a JPEG and add that instead.
A PST fileDoes not readSureRedact says so: export the individual messages you need — as .msg or .eml — and add those instead.
OneNote sections (.one)Does not readSureRedact says so, and says what to do: in OneNote, open the notebook and use File → Export to save the section as a PDF, then add that PDF. The detail is in Getting OneNote notebooks into SureRedact.
A OneNote index (.onetoc2)Nothing to readNothing — it is the notebook’s table of contents, not its pages. SureRedact says so.
Microsoft Forms and similar app dataReads the HTML if Purview converted it; does not read a .form fileOpen the form in Microsoft Forms and export its responses to Excel, which SureRedact converts
Audio and video (.m4a, .mp3, recordings)Does not readSureRedact says so: audio is not redacted here, and can hold personal data that still has to be dealt with outside the product. Listen, and decide, outside it.
A zip file inside the exportDoes not readUnzip it into the folder first

Email

SureRedact reads Outlook .msg files — the form Purview exports individual messages in — as the emails they are: the headers, the body, and every attachment, each converted as if it were a file in the folder. A message attached to a message is converted as a message too, several levels deep. Email conversion is in both editions from 1.2.0; Office conversion is Professional.

Three things it tells you rather than guesses at. Each is a note on the converted document, where you will see it:

  • A body Outlook stored only as compressed rich text is not read. Outlook has saved a plain-text and an HTML copy of the body alongside for twenty years, so this is rare. When it happens the headers and attachments are included, the body is not, and the note says open the original in Outlook to read it. That is the one case where the per-message route below is still needed.
  • A message with only an HTML body has its text taken from the HTML; layout and pictures are not shown.
  • An attachment with no content, or an attached message that could not be read, is named in a note and left out — never dropped in silence.

Bcc is kept. On a message your organisation sent, who else received it is part of the record, and it is disclosable.

The per-message route, if you need it — for a rich-text-only body, or for a message you hold in Outlook rather than in the export:

  • New Outlook for Windows, or Outlook on the web. Open the message; More actions › Save as (new Outlook) or More actions › Download (web). The .eml file that results goes into the Matter folder, attachments with it.
  • Classic Outlook for Windows. File › Print, printer Microsoft Print to PDF. Attachments do not come with the PDF — File › Save Attachments into the same folder.
  • Outlook for Mac. Drag the message into a Finder folder for an .eml. Some recent versions have broken this; if dragging does nothing, print to PDF as above.

Two things worth knowing

  • The friendly-name option puts subject lines in filenames. A filename travels into the pack as the document’s name. If a subject line says more than it should, rename the file before you import it.
  • A message’s attachments are part of the message. When you keep a message and withhold its attachment, or the other way round, say why in the reason box — the requester reads that.

Sources

Related guides