nope.
after SE generated 9 records for the content, the correspondent datamap (pointing to that 9 records) got self_encrypted again to generate 3 records (for the datamap) and return with a datamap pointing to that 3 datamap records
3 Likes
So when he said self-encryption will always return 3 chunks + datamap is only for datamaps? It was the "always" that got me. Usually always means always and not just for some things 
Continuing on then for a 1000 record file will there be multiple levels of datamaps till it reduces to just one? Or?
2 Likes
datamap is a list of ChunkInfo, whose size is around 80 Bytes each.
a 1000 record file will have a datamap with size of 80k, which is way enough for 3 chunks to hold it.
In theory, only file composed by over 153k records (about 600GB large) will have the multiple layer datamap to be shrinked recursively
7 Likes
I think he means: the returned datamap always points to 3 chunks.
1 Like
And you just said one datamap can point to 1000 records. So there is some contradiction there 
all fine, you answered the question. And there doesnāt seem too much change in the concept, just more in the data_map record type
2 Likes
that datamap is the inner one, not the final one returned from SE 
to be clarify, you can consider this as file datamap (which points to content records), meanwhile the returned datamap as root datamap.
and root_datamap always points to 3 records, meanwhile file_datamap points to X records depends on file_size
For file size <= 12 MB (i.e. <= 3 records) returned root_datamap is file_datamap itself
For file size > 12 MB (i.e. > 3 records) returned root_datamap is NOT file_datamap
3 Likes
Ah OK, and now I see why people were thinking 4 extra records.
So a file over 3 records and below ~153K records requires a āfile_datamapā that is being stored along with the file (ie file SE records + 3) and then there is 3 records for the root_datamap.
The root datamap is stored on autonomi if public.
and over ~153K records has another level of data maps pointing to the file_datamaps.
3 Likes
thanks to you @neo and @qi_ma for digging and clarifying =)
ps: the breaking change in datamaps did catch @aatonnomicc and me with friends too (even though we knew it existed) when we sent around media in friends ā¦ because I was on my custom branch of dweb with new autonomi client with correct fork detection/resolution ā¦ and he was with the official dweb ā¦ took us a while to think of this when he couldnāt receive my files but I was able to receive his ā¦
7 Likes
it was late at night though 
4 Likes
true true
just thought Iād mention such a breaking change in the most fundamental datatype is a nasty move that hits hard on many occasions xD
4 Likes