Understanding BSON Notation

Question

I was trying to understand BSON Notations from the site BSON Site. However I was unable to understand the reason behind the correlations.

I also referred the following questions, but am not convinced because of the following reasons.

Question 1 : Not familiar with ruby implementations

Question 2 : I understood the byte allocation. But unsure about the notations.

I would like to know how the bson object is formed for the following examples in BSON Site

1.{"hello": "world"}
2.{"BSON": ["awesome", 5.05, 1986]}

Answer 1

{"hello": "world"}

\x16\x00\x00\x00
  \x02 hello\x00 \x06\x00\x00\x00 world\x00
\x00

(overall: 22 bytes)

The first four bytes contain the overall length as a 32-bit little-endian integer.

\x16\x00\x00\x00 => That's 22 in decimal.

Now comes the first element. The first byte gives the type of data.

\x02 => That's a UTF-8 string.

Then comes the name of the first element, as a null-terminated string.

hello\x00

Next comes the data of the element in the previously given type, in this case a string. For scannability (so you can quickly skip when you don't need them), strings start with their length, and are null-terminated.

\x06\x00\x00\x00 => That's length 6.

world\x00

Now would come subsequent elements, if there were any. The whole thing is terminated with a null byte.

\x00

---

{"BSON": ["awesome", 5.05, 1986]}

\x31\x00\x00\x00
  \x04 BSON\x00 \x26\x00\x00\x00
                  \x02 0\x00 \x08\x00\x00\x00 awesome\x00
                  \x01 1\x00 \x33\x33\x33\x33\x33\x33\x14\x40
                  \x10 2\x00 \xc2\x07\x00\x00
                \x00
\x00

(overall: 49 bytes, array: 38 bytes)

The first four bytes contain the overall length as a 32-bit little-endian integer.

\x31\x00\x00\x00 => That's 49 in decimal.

Now comes the first element. The first byte gives the type of data.

\x04 => That's an array.

Then comes the name of the first element, as a null-terminated string.

BSON\x00

Next comes the data of the element in the previously given type, in this case an array.

[Quote: "The document for an array is a normal BSON document with integers for the keys, starting with 0 (..)"]

For scannability and because they form document in their own right, arrays start with their length, and are null-terminated.

\x26\x00\x00\x00 => That's 38 in decimal.

Now comes the first element of the array. The first byte gives the type of data.

\x02 => That's a UTF-8 string.

Then comes the name of the first element of the array, null-terminated.

0\x00 => That's key 0.

Next comes the data of the element in the previously given type, in this case a string. Strings start with their length, and are null-terminated.

\x08\x00\x00\x00 => length 8

awesome\x00

Now comes the second element of the array. The first byte gives the type of data.

\x01 => That's a double floating point number.

Then comes the name of the second element of the array, null-terminated.

1\x00 => That's key 1.

Next comes the data of the element in the previously given type, in this case a double floating point number.

\x33\x33\x33\x33\x33\x33\x14\x40 => That's 5.5.

Now comes the third element of the array. The first byte gives the type of data.

\x10 => That's a 32-bit integer.

Then comes the name of the third element of the array, null-terminated.

2\x00 => That's key 2.

Next comes the data of the element in the previously given type, in this case a 32-bit integer.

\xc2\x07\x00\x00 => That's 1986.

The array is terminated with a null byte.

\x00

The whole thing is terminated with a null byte.

\x00