The customer is a fashion company which has been based in Bavaria since the 1970th. It operates over 400 retail spaces (as of September 2018) in Germany, Austria, Belgium, Norway, the UK, Switzerland and the Netherlands, as well as operating its own online shops in Germany, Austria, Switzerland, the UK and France. The company's design team creates the trendy, high-quality womenswear collections which are supplemented by accessories and shoes.
Like many other companies, the fashion brand was faced with the great challenge of taking an eCommerce system based on Magento 1, which had grown over the years and is already working very well, to a new level and thus creating the basis for further growth and corresponding future security. With the migration from Magento 1 to Magento 2, which is associated with a multitude of innovations and technical improvements, the fashion brand customers will be offered a new shopping experience.
The biggest challenge by far was the data import/export from and to TB.One as the fashion brand's central data platform as well as the data migration from Magento 1 to Magento 2. TB.One offers for each of the 6 available websites separate XML files with different XML structures as well as different option- and category IDs. These files must be merged and extracted into separate CSV files in the standard M2IF CSV formats for attributes and attribute values, as well as products and product pricing. In addition to the permanent import process, which runs once a day, it should also be possible to manually trigger the import for individual websites. In addition, product prices and inventory information from a separate TB.One feed, which also provides XML files, must be processed at intervals of 60 and 15 minutes.
In order to improve performance and optimize indexing processes, both full and delta import modes are supported for product, warehouse and price imports. The delta mode for updating inventory and price information is used during the day. In all cases, the XML contains a time stamp to resume the delta import from the last time stamp. The implemented TB.One connector has been optimized for powerful, storage-efficient XML processing that merges and enriches XML files with category mappings. Data maintenance takes place at the Store View level. The data is extracted into the standard CSV target files required for the attribute/attribute option and product import. The so-called process pipelines streamline the import processes and ensure data consistency between the import steps. In addition to the ongoing nightly import, it is also possible to manually trigger the import for a specific website by simply selecting the appropriate pipeline in the admin backend. The indexing process for frequent price and inventory imports is also mapped via the process pipelines to avoid the entire cache becoming invalid after an import.
The processing of the 6 XML files with several hundred thousand lines per file takes about 26 minutes and requires less than 1 GB storage space. The import process that starts after the conversion process takes approximately 10 minutes to import the attributes and products into Magento. Finally, another step updates the product status, which takes another 10 minutes. The main import is usually done once a night and therefore NOT influences the daily business and the performance of the shop during the main shopping hours. The Single Transaction Mode ensures that product, price and inventory updates, whether full or delta, are always consistent and either complete or not at all. In addition, Single Transaction mode can greatly speed up the import process by processing all DML statements at once to obtain the data in the database as soon as the transaction is transferred.
|EIP 1.0||Product Batch|
|Performance Booster (necessary for Single Transaction Mode when using EE)|
|M2IF 1.0||Single Transaction Mode|
|Import frequency||Stock [Frequency: every 30 minutes, Runtime: ~3 seconds]|
|Attribute [Frequency: once per day, Runtime: ~6 seconds]|
|Product + Prices [Frequency: once per day, Runtime: ~10 minutes]|
|Magento Version||Magento EE 2.1|