Spatial Database Engine inquiry self http://www.goldenbg.com/article.asp? Id = 819)

Place GIS data in RDBMS, however, generally RDBMS does not provide the GIS data type (such as the point, line, polygon, And the topological relationship between these feature and projection coordinates ), RDBMS only supports a small amount of data types: int, float, double, blog, long, Char, etc., generally numbers, strings and binary data. In addition, RDBMS does not extract the storage of GIS data types, nor does it provide operations on these basic types (for example, determining the inclusion relationship, adjacent, intersection, difference, distance, shortest path, etc ).

To store GIS data types and support operations on GIS data types. We can use twists and turns to achieve these goals.

1. Storage: store featrue in the form of a binary field, and create a description of the binary data (called spatial metadata ).

2. Operation: Create a function package for each feature.

A. Read the spatial metadata first, and fill the binary data of featrue into the data part of the featrue object according to the metadata.

B. When writing data, write the spatial metadata, serialize the data part of featrue to binary data, and store the data in the feature field.

C. Spatial operation: the content of spatial operation is the data part of the fatrue object, and spatial operation is a member function of the fatrue class. Or create a function package for each feature separately.

Next we will first explore the storage of spatial data types. Storage mainly involves the metadata of GIS data. We need to manage various spatial data.

Featrue of the same type is classified as featrue class. Each feature class corresponds to a binary field in the database, and each feature object corresponds to a specific value of this binary field. In order to store the attribute data and spatial data of the feature in GIS and manage them in a unified manner. We have established these concepts.

Dataset: corresponds to a table in the database. The table contains the binary field of featrue and a general attribute field. It also includes the metadata fields of each feature object, such as perimeter, area, and auxiliary fields.

Data source: corresponds to a group of tables in the database, which consists of two parts: one is a dataset, and the other is a table (called a space metadata table) that describes the dataset ). It is usually a set of spatial metadata tables under each schma (or database.

Auxiliary database objects: including indexes, triggers, sequences ......

The metadata table includes:

A: Describe the table of each dataset. The fields in this table include the surrounded box of each dataset. And the index type and name created for the dataset.

B: Table that describes the dataset fields. The fields in this table include: DataSet name, field name, field type, field alias, dataset alias, and dataset type ......

C: Metadata table of the Secondary object: including the table that describes the relationship between the index and the dataset, including the table that describes the relationship between the trigger and the dataset, including the table that describes the relationship between the sequence and the dataset. A table that describes the relationship between datasets (such as a network dataset, which is composed of a point-type dataset and a line-type dataset. This describes the topological relationship between a point and a line)

We will only discuss the storage of Vector GIS data. Next we will explore the storage of grid gis data.

There are also raster datasets, but the field type has changed slightly. In order to accelerate data access, we should allow GIS operations to only access the data it requires. Taking a remote sensing image as an example, we can cut a large image into a series of small blocks according to certain rules, then, store these small image blocks in binary form to the image field, and at the same time, the offset location of the small image block (with the actual location, and the offset location and the size of each graph block will know the geographical range of the graph block.) stored as the attribute field of the table. When retrieving data in the future, you need to first check the metadata content of the graph block to obtain the geographical range (or simply store the geographical range of each graph block ), in this way, you will know which records need to be retrieved during display, and then retrieve the image data of the records to be displayed, you can splice the data according to the geographical range (the display and operation of vector data is also the same ). It is important to define the block size rules, which greatly affects the speed.

Generally, a band corresponds to a grid dataset. When displaying images, you need to read multiple grid datasets at the same time and splice them together during display.

We also need to describe the metadata tables of these raster datasets. For example, these raster datasets correspond to the same image file, the geographical coordinates, and the image pyramid information.

Next we will discuss the image pyramid. In order to speed up data display, some spatial database engines dilute the most primitive data and use each layer of gold tower as a grid dataset. Generally, the power series is used. This is a way to change the space for time.

I have never studied image compression technology. Wait a while and let's talk about it.

With these metadata tables and feature tables, This laid the foundation for GIS operations.