def binary_search_table_interior(fd, cell_pointer, target_value,
right_most_pointer):
global table_interior_page_read
table_interior_page_read += 1
n = len(cell_pointer) // 2
cell = cell_pointer[n]
fd.seek(cell)
left_pointer = util.read_to_int(fd, 4)
row_id = util.decode_varint(fd)[0]
if len(cell_pointer) == 1:
if target_value <= row_id:
return left_pointer
else:
return right_most_pointer
if len(cell_pointer) == 2:
if target_value <= row_id:
cell_0 = cell_pointer[0]
fd.seek(cell_0)
left_pointer_0 = util.read_to_int(fd, 4)
row_id_0 = util.decode_varint(fd)[0]
if target_value <= row_id_0:
return left_pointer_0
return left_pointer
else:
return right_most_pointer
else:
if target_value <= row_id:
return binary_search_table_interior(fd, cell_pointer[:n + 1],
target_value,
right_most_pointer)
else:
return binary_search_table_interior(fd, cell_pointer[n + 1:],
target_value,
right_most_pointer)
def binary_search_table_leaf(fd, cell_pointer, target_value):
global data_page_read
data_page_read += 1
n = len(cell_pointer) // 2
cell = cell_pointer[n]
fd.seek(cell)
payload_byte = util.decode_varint(fd)
row_id, row_id_size = util.decode_varint(fd)
if n == 0:
if target_value == row_id:
header_size = util.decode_varint(fd)[0]
fd.seek(cell + payload_byte[1] + row_id_size + header_size)
emp_id = util.read_to_int(fd, 3)
record = fd.read(payload_byte[0] - 3)
employee = util.construct_record(emp_id, record)
return employee
else:
print('row id does not exist in this page, something is wrong.')
exit(1)
else:
if target_value < row_id:
return binary_search_table_leaf(fd, cell_pointer[:n], target_value)
else:
return binary_search_table_leaf(fd, cell_pointer[n:], target_value)
def read_table_leaf_cell(fd, usable_size, cell):
fd.seek(cell)
payload_byte = util.decode_varint(fd)
rowid_byte = util.decode_varint(fd)
header_size = util.decode_varint(fd)[0]
_, initial_payload_size = util.overflow(13, payload_byte[0], usable_size)
fd.seek(cell + payload_byte[1] + rowid_byte[1] + header_size)
emp_id = util.read_to_int(fd, 3)
reset_of_record = fd.read(initial_payload_size - 3)
# if overflow:
# TODO
# pass
employee = util.construct_record(emp_id, reset_of_record)
return employee
def search_index_leaf_range(fd, cell_pointer, lower_bound, upper_bound):
global index_leaf_page_read
index_leaf_page_read += 1
result = []
for cell in cell_pointer:
fd.seek(cell)
payload_byte = util.decode_varint(fd)
header_size = util.decode_varint(fd)[0]
emp_id_size = util.read_to_int(fd, 1)
row_id_size = util.read_to_int(fd, 1)
fd.seek(cell + payload_byte[1] + header_size)
emp_id = util.read_to_int(fd, emp_id_size)
row_id = util.read_to_int(fd, row_id_size)
if emp_id > lower_bound and emp_id < upper_bound:
result.append(row_id)
return result
def binary_search_index_interior(fd, cell_pointer, target_value,
right_most_pointer):
global index_interior_page_read
index_interior_page_read += 1
n = len(cell_pointer) // 2
cell = cell_pointer[n]
fd.seek(cell)
left_pointer = util.read_to_int(fd, 4)
key_payload_byte = util.decode_varint(fd)
key_header_size = util.decode_varint(fd)[0]
emp_id_size = util.read_to_int(fd, 1)
fd.seek(cell + 4 + key_payload_byte[1] + key_header_size)
emp_id = util.read_to_int(fd, emp_id_size)
if len(cell_pointer) == 1:
if target_value < emp_id:
return left_pointer
else:
return right_most_pointer
elif len(cell_pointer) == 2:
if target_value < emp_id:
cell_0 = cell_pointer[0]
fd.seek(cell_0)
left_pointer_0 = util.read_to_int(fd, 4)
key_payload_byte_0 = util.decode_varint(fd)
key_header_size_0 = util.decode_varint(fd)[0]
emp_id_size_0 = util.read_to_int(fd, 1)
fd.seek(cell_0 + key_payload_byte_0[1] + key_header_size_0)
emp_id_0 = util.read_to_int(fd, emp_id_size_0)
if target_value < emp_id_0:
return left_pointer_0
return left_pointer
else:
return right_most_pointer
else:
if target_value < emp_id:
return binary_search_index_interior(fd, cell_pointer[:n + 1],
target_value,
right_most_pointer)
else:
return binary_search_index_interior(fd, cell_pointer[n + 1:],
target_value,
right_most_pointer)
def search_index_interior_range(fd, cell_pointer, lower_bound, upper_bound,
right_most_pointer):
global index_interior_page_read
index_interior_page_read += 1
result = []
for cell in cell_pointer:
fd.seek(cell)
left_pointer = util.read_to_int(fd, 4)
key_payload_byte = util.decode_varint(fd)
key_header_size = util.decode_varint(fd)[0]
emp_id_size = util.read_to_int(fd, 1)
fd.seek(cell + 4 + key_payload_byte[1] + key_header_size)
emp_id = util.read_to_int(fd, emp_id_size)
if emp_id > lower_bound and emp_id < upper_bound:
result.append(left_pointer)
if emp_id > upper_bound:
result.append(left_pointer)
break
if emp_id < upper_bound:
result.append(right_most_pointer)
return result
def binary_search_index_leaf(fd, cell_pointer, target_value):
global index_leaf_page_read
index_leaf_page_read += 1
n = len(cell_pointer) // 2
cell = cell_pointer[n]
fd.seek(cell)
payload_byte = util.decode_varint(fd)
header_size = util.decode_varint(fd)[0]
emp_id_size = util.read_to_int(fd, 1)
row_id_size = util.read_to_int(fd, 1)
fd.seek(cell + payload_byte[1] + header_size)
emp_id = util.read_to_int(fd, emp_id_size)
row_id = util.read_to_int(fd, row_id_size)
if n == 0:
if target_value == emp_id:
return row_id
else:
return None
else:
if target_value < emp_id:
return binary_search_index_leaf(fd, cell_pointer[:n], target_value)
else:
return binary_search_index_leaf(fd, cell_pointer[n:], target_value)