Improve compatibility with csvsql

[iainelder]

Using aws-org-inventory version 0.2.0.

cases.csv was created using this command:

poetry run aws-org-inventory support describe_cases cases > cases.csv

$ csvsql --snifflimit 0 --query 'SELECT 1 FROM cases' -- cases.csv
/home/isme/.local/pipx/venvs/csvkit/lib/python3.8/site-packages/agate/utils.py:276: UnnamedColumnWarning: Column 0 has no name. Using "a".
OperationalError: (sqlite3.OperationalError) duplicate column name: status
[SQL: 
CREATE TABLE cases (
	a FLOAT, 
	"Id" FLOAT, 
	"Arn" VARCHAR, 
	"Email" VARCHAR, 
	"Name" VARCHAR, 
	"Status" VARCHAR, 
	"AssumeRoleSuccess" BOOLEAN, 
	"caseId" VARCHAR, 
	"displayId" FLOAT, 
	subject VARCHAR, 
	status VARCHAR, 
	"serviceCode" VARCHAR, 
	"categoryCode" VARCHAR, 
	"severityCode" VARCHAR, 
	"submittedBy" VARCHAR, 
	"timeCreated" TIMESTAMP, 
	"recentCommunications" VARCHAR, 
	"ccEmailAddresses" VARCHAR, 
	language VARCHAR, 
	CHECK ("AssumeRoleSuccess" IN (0, 1))
)

]
(Background on this error at: http://sqlalche.me/e/13/e3q8)

Several issues to fix here (maybe split them out):

• Avoid duplicate column names by removing less-useful columns added by botocove (Arn, Email, Name, Status) so that only Id remains
• Rename Id column to InventoryAccountId (not just AccountId because it may still clash with stack set APIs)
• Remove or replace the unnamed index column (this forced me to use snifflimit)
  • replace with AccountId and Id of resource (may be hard to determine resource Id in the general case)
  • can you have a data frame without an index, or at least hide it on export?
  • replace with a monotonically increasing id