When a command's output, a search or a fetched page produces more text than a chat carries, DataMind Oda shortens the result and adds a short note; the largest results are also written in full to a file on your computer. This page explains the notes you see, where the full text goes, and when these files are cleared up.
Oda keeps results within a size a chat can carry. When more arrives than fits, the result you see is shortened, and a note counts what was left out:
Omitted 12 bytes. In a text result, the number is the size of the part left out of the chat.Omitted 3 items. In a list of findings, the number is how many entries were left out.More lines were omitted. instead of a number; when nothing was left out, there is no note.Omitted 25 items. Results capped at 100. Narrow the pattern, path, or include to see more.(Omitted N bytes. Full formatted result stored at: <locator>. Use read with offset/ limit, or grep this path to search within it. )The part that stays depends on the kind of result:
| Kind of result | The shortened form keeps |
|---|---|
| A file search or a text search | The first matches |
| A web search | The first sources |
| A fetched page | The start of the page |
| A command's output | The end of the output, or both its start and its end |
Storing is best effort:
The result held more findings than the chat keeps, so Oda showed the part that fits and counted the rest. Nothing was thrown away: where the full text can be kept, Oda writes it to a file on your computer, and the note names it. Ask Oda to read the file or to search inside it when you need the rest.
Items are counted one by one when the result is a list of findings, such as search matches or sources. Text is measured in bytes, so a cut text reports the size of the part left out, and a list reports how many entries were left out. When the exact number is not known, the note says More lines were omitted. without a count.
It depends on the kind of result: searches and fetched pages keep their start, and a command's output keeps its end, or both its start and its end. The table above lists each kind. Cuts fall between characters, so you never see half a symbol.
No. A truncated note says a size limit was reached, not that the source was incomplete or the step failed; failures are reported as failures, in their own words. What a cut left out is counted in its note, and the fuller text is kept where the note says.
By default, into a private folder under your computer's temporary files area, created the first time a file is needed; your installation can choose a different folder. Inside it, each chat has a folder of its own, and the files can be read only by your user account. Nothing inside Oda browses these files; the note in the chat is the usual pointer.
Files older than 30 days are removed when Oda starts; your installation can choose a different period or turn the clearing off. The clearing runs at start-up only, so an Oda that stays open for a long time clears its older files at the next start. Files are kept on purpose: a chat you resume or fork can still read a result it stored earlier until the file ages out. Nothing in Oda deletes a stored file on request.
Storing is best effort, and two cases leave the result as it was: Oda could not write the file, or no note small enough for the size set could be made. In both cases the result stays in the chat and the step does not fail; a failed save never turns a good result into a lost one.
Yes, until the file ages out. A retry, a resumed chat and a forked chat still point at the same file, and Oda reads it back in parts or searches inside it when you ask. Once the file has been cleared, the note has nothing behind it; ask Oda for the work again if you need the text.