We joined 100,000 card transactions to a customer segment table to compute spend by segment, but only 85,000 matched and 15,000 came back with no segment. Spot checks show the customer ID looks like 00123 in the transaction table but 123 in the segment table. How would you investigate and fix this, and what else could cause unmatched records?

我们把10万笔信用卡交易关联到客户分群表上算各群的消费，结果只有8.5万笔匹配上，有1.5万笔的分群是空的。抽查发现交易表里的客户号像00123，分群表里是123。你怎么排查和修复？还有哪些原因会导致匹配不上？