You said it is only 1000 rows every 2 hours in the comments. That is a acceptable amount of data for a C# programm. I would say the big question left is wich output format you use.
.CSV is the simplest one. This format can be done with a File.WriteLine() and some string concaction. There is no build in CSV parser or writer code I am aware off in C#, but there is plenty of 3rd party code.
.XLS requires the (t)rusty Office COM Interop. That requires office to be installed and does not work from a non-interactive session (like a Windows Service). On top of all the normal issues for using COM interop.
There is the odd "export to XLS" function on existing classses, but those are rare, far inbetween and about everything you get. Unfortunately as we always had COM Interop as fallback, we never quite developed a standalone library for working with .XLS. Ironically working with this old format is harder from C#/.NET then it would be from Java.
.XLSX however is easier. It can be written using the OpenXML SDK. Or the XML writer and ZipArchive class: At their core all the ???x formats are a bunch of .XML files in a renamed .ZIP container. There should even be 3rd party code out there to make using the SDK easier.
.CSV is the lowest common denominator and propably the easiest to create. However if a user is supposed to open this document, the lack for formating might become an issue.
.XSLX would be my choice if you need a user to open it.
.XSL I would avoid like a swarm of angry bees.